{"auto_groups":["CodexTeam","CodexPro","ClaudeKiro","ClaudeMax","default","Gemini","Grok","NIM"],"data":[{"model_name":"nvidia/ising-calibration-1.5-31b","description":"Ising Calibration 1.5 is a vision-language model for quantum-computing calibration workflows. It interprets calibration experiment plots and produces technical analysis to support device tuning.","tags":"Vision,Reasoning","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"],"pricing_version":"5a90f2b86c08bd983a9a2e6d66c255f4eaef9c4bc934386d2b6ae84ef0ff1f1f"},{"model_name":"gpt-6-astra","description":"GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":2,"enable_groups":["CodexCyber","CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 10 + c * 50 + cr * 1 + cc * 12.5) : tier(\"long_context\", p * 20 + c * 75 + cr * 2 + cc * 25)"},{"model_name":"go/deepseek-v4-flash-0731","description":"DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.","tags":"Vision,Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5 \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.32 + c * 1.28 + cr * 0.0064)\n  : tier(\"off_peak\", p * 0.16 + c * 0.64 + cr * 0.0032)"},{"model_name":"go/ali/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...","icon":"Zhipu.Color","tags":"Reasoning,Tools,Vision,Video,1M","vendor_id":12,"quota_type":0,"model_ratio":0.18,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.2875,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-5.6-luna","description":"GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":2,"quota_type":0,"model_ratio":0.1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 0.2 + c * 1.2 + cr * 0.02 + cc * 0.25) : tier(\"long_context\", p * 0.4 + c * 1.8 + cr * 0.04 + cc * 0.5)"},{"model_name":"oc/longcat-2.5-preview","description":"LongCat 2.5 Preview is a preview endpoint in the LongCat family for conversational assistance, text generation, and coding tasks. Features and behavior may change during the preview.","tags":"Reasoning,Tools","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"oc/minimax-m3.1-flash","description":"MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":6,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"sensevoicesmall","description":"SenseVoiceSmall is a speech-understanding model for audio transcription and related audio analysis. It can be used to convert spoken content into text for search, accessibility, and downstream processing.","tags":"Audio","vendor_id":99,"quota_type":0,"model_ratio":0.000075,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/nemotron-3.5-super-vl-120b-a12b","description":"NVIDIA Nemotron 3.5 Super VL is an open-weight vision-language model with 120B total and 12B active parameters. It supports reasoning, coding, tool use, document understanding, and video understanding, with text, image, and video inputs and text output.","icon":"Nvidia.Color","tags":"Reasoning,Tools,Vision,Video,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-nano-banana-2.1","description":"Nano Banana 2.1 (Gemini Nano Banana 2.1) is Google's image generation and editing model on the Flash tier, succeeding Nano Banana 2 and Nano Banana Pro. It improves product recontextualization, mask- and ink-based editing, and factual accuracy, and renders photorealistic skin tones, detailed materials, lighting, and coherent backgrounds. It accepts text and image inputs, returns images with optional text, and supports 1K, 2K, and 4K output plus extended aspect ratios via the [image_config API Parameter](https://openrouter.ai/docs/features/multimodal/image-generation#image-aspect-ratio-configuration).","icon":"Gemini.Color","tags":"Image,Vision,Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.035,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"gpt-image-2.5-sunburst","description":"GPT Image 2.5 Sunburst is an image generation and editing model focused on precision-oriented visual work. It can be used for detailed compositions, controlled edits, and production assets.","tags":"Image,Reasoning,Tools","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.075,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"kimi-k2.8-preview","description":"Kimi K2.8 Preview is a preview endpoint in the Kimi family for conversational assistance, drafting, summarization, and text-based problem solving. Features and behavior may change during the preview.","tags":"Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":0.2925,"model_price":0,"owner_by":"","completion_ratio":4.153846153846154,"cache_ratio":0.26153846153846155,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"kimi-k2.6","description":"Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration.","tags":"Reasoning,Tools,Files,Open Weights,Vision,262.1K","vendor_id":9,"quota_type":0,"model_ratio":0.4875,"model_price":0,"owner_by":"","completion_ratio":4.153846153846,"cache_ratio":0.2,"create_cache_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-6-luna","description":"GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic tasks, and at higher reasoning effort it can take on complex software engineering and computer-use tasks that previously called for a Sol-tier model. It shares the GPT-6 family's gains in factual reliability and its clearer, more concise communication style.","icon":"OpenAI","tags":"Vision,Files,Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":2,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 0.1 + cr * 0.01 + cc * 0.125 + c * 0.5) : tier(\"long_context\", p * 0.2 + cr * 0.02 + cc * 0.25 + c * 0.75)"},{"model_name":"go/hy3","description":"Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...","icon":"Hunyuan.Color","tags":"Reasoning,Tools","vendor_id":69,"quota_type":0,"model_ratio":0.08,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"claude-sonnet-4-6","description":"Claude Sonnet 4.6 is an Anthropic model for coding, agents, and professional work, including iterative development and complex codebase navigation.","icon":"Claude.Color","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":1.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.6,"enable_groups":["ClaudeR1","ClaudeR4","ClaudeR2","ClaudeR3","ClaudeKiro"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 3 + c * 15 + cr * 0.3 + cc * 3.75 + cc1h * 6)"},{"model_name":"agnes-image-2.0-flash","description":"Agnes Image 2.0 Flash is an image-generation endpoint for creating visual content from prompts. It can be used for illustrations, design concepts, and creative assets.","tags":"Image","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.000004,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai-video"]},{"model_name":"go/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Vision,Reasoning,Tools","vendor_id":12,"quota_type":0,"model_ratio":0.064,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.2875,"create_cache_ratio":0.000007,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/glm-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.","tags":"Reasoning,Tools","vendor_id":12,"quota_type":0,"model_ratio":0.64,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"selfdeploy/deepseek-v4-flash","description":"DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.","tags":"Reasoning,Tools,1M","vendor_id":7,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"bge-m3","description":"BGE-M3 is BAAI's multilingual embedding model for dense, sparse, and multi-vector retrieval. It supports more than 100 languages and inputs up to 8,192 tokens.","tags":"Embeddings","vendor_id":32,"quota_type":0,"model_ratio":0.015,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.8-flash","description":"Gemini 3.8 Flash is a Google model for software engineering, agentic tasks, and multi-step reasoning, building on Gemini 3.7 Flash.","tags":"Reasoning,Tools","vendor_id":8,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["Gemini","GeminiUltra"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"gemini-3-flash","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,1M","vendor_id":8,"quota_type":0,"model_ratio":0.25,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["Gemini","GeminiUltra"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"nvidia/nemotron-3-embed-1b","description":"Nemotron 3 Embed 1B converts text into vector representations for semantic retrieval. It can be used for enterprise search, document matching, and retrieval-augmented generation.","tags":"Embeddings","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"m/gpt-image-2","description":"GPT Image 2 is an image generation and editing model for creating illustrations, visual designs, and image transformations from text prompts and reference images.","icon":"OpenAI","tags":"Image,Vision","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"oc/nemotron-3-ultra","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE).","tags":"Reasoning,Tools,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/grok-4.6","description":"Grok 4.6 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM. It is succeeded by [Grok 4.7](/x-ai/grok-4.7).","icon":"XAI","tags":"Vision,Files,Reasoning,Tools","vendor_id":5,"quota_type":0,"model_ratio":0.1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 0.2 + c * 0.6 + cr * 0.05) : tier(\"long_context\", p * 0.4 + c * 1.2 + cr * 0.1)"},{"model_name":"MiniMax-M3","description":"MiniMax-M3 is a multimodal foundation model from MiniMax that accepts text, image, and video inputs and produces text, with support for long-horizon agentic work and coding.","tags":"Reasoning,Tools,Files,Open Weights,Vision,1M","vendor_id":6,"quota_type":0,"model_ratio":0.0945,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"kimi-k2.7-code","description":"MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts.","tags":"Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":0.4875,"model_price":0,"owner_by":"","completion_ratio":4.153846153846,"cache_ratio":0.2,"create_cache_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/deepseek-v4-pro-0813","description":"DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5 \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 1.44000 + c * 4.32000 + cr * 0.048000)\n  : tier(\"off_peak\", p * 0.72000 + c * 2.16000 + cr * 0.024000)"},{"model_name":"z-ai/glm-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.","tags":"Reasoning,Tools","vendor_id":12,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-5.6-terra","description":"GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":2,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2 + cc * 2.5) : tier(\"long_context\", p * 5 + c * 22.5 + cr * 0.5 + cc * 6.25)"},{"model_name":"anti/gemini-3.1-flash-lite","description":"Gemini 3.1 Flash Lite is Google’s GA high-efficiency multimodal model optimized for low-latency, high-volume workloads.","tags":"Vision,Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.01,"owner_by":"","completion_ratio":0,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"unreal-speech-v8","description":"Unreal Speech V8 is a text-to-speech endpoint for turning written content into spoken audio. It can be used for narration, reading assistance, and voice-enabled applications.","tags":"Audio","vendor_id":99,"quota_type":0,"model_ratio":0.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"MiniMax-M2.7","description":"MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.","tags":"Reasoning,Tools,Open Weights,204.8K","vendor_id":6,"quota_type":0,"model_ratio":0.1575,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-haiku-4-5-20251001","description":"Claude Haiku 4.5 is an Anthropic model designed for low-latency, efficient inference.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":1,"quota_type":0,"model_ratio":0.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["ClaudeR3","ClaudeKiro","ClaudeMax","ClaudeR1","ClaudeR4","ClaudeR2"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 1 + c * 5 + cr * 0.1 + cc * 1.25 + cc1h * 2)"},{"model_name":"glm-z1-9b-0414","description":"GLM Z1 9B 0414 is a reasoning-oriented language model for multi-step analysis, coding, and problem solving. It is suited to tasks that benefit from additional reasoning before an answer.","tags":"32K","vendor_id":12,"quota_type":0,"model_ratio":0.003,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4-pro-0813","description":"DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":0.2025,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5\n  \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12)\n    || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.810 + c * 2.430 + cr * 0.0270)\n  : tier(\"off_peak\", p * 0.4050 + c * 1.2150 + cr * 0.01350)"},{"model_name":"nvidia/nemotron-3-super-120b-a12b","description":"NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.","tags":"Reasoning,Tools,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"go/qwen3.7-max","description":"Qwen3.7-Max is a text model in Alibaba's Qwen3.7 series, designed for agent-centric workloads including coding, office work, and productivity tasks.","tags":"Files,Reasoning,Tools","vendor_id":11,"quota_type":0,"model_ratio":0.96,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"qwen2.5-7b-instruct","description":"Qwen2.5 7B Instruct is an instruction-tuned model in the Qwen2.5 family, with capabilities in knowledge tasks and coding.","tags":"Tools,33K","vendor_id":11,"quota_type":0,"model_ratio":0.0015,"model_price":0,"owner_by":"","completion_ratio":3.3333333333333335,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"stepfun-ai/step-5-preview","quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"anti/gemini-3-flash","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,1M","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0175,"owner_by":"","completion_ratio":0,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"gemini-3.1-flash-image","description":"Gemini 3.1 Flash Image, also known as Nano Banana 2, is Google's model for image generation and image editing.","tags":"Reasoning,Files,Vision,65.5K","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.04,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"agnes-2.5-flash","description":"Agnes 2.5 Flash is a text-generation endpoint for conversational assistance and written content. It can be used for question answering, drafting, summarization, and text transformation.","tags":"Text","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.000004,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai-video"]},{"model_name":"qwen3.8-flash","description":"Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.","tags":"Vision,Files,Reasoning,Tools","vendor_id":11,"quota_type":0,"model_ratio":0.06,"model_price":0,"owner_by":"","completion_ratio":3.375,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"mimo-v2.5-pro","description":"MiMo-V2.5-Pro is a Xiaomi model for agentic applications, complex software engineering, and long-horizon tasks.","icon":"XiaomiMiMo","tags":"Reasoning,Tools,Open Weights,1M","vendor_id":16,"quota_type":0,"model_ratio":0.2325,"model_price":0,"owner_by":"","completion_ratio":1.967741935483871,"cache_ratio":0.03225806451612903,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"m/nano-banana-2","description":"Nano Banana 2, also known as Gemini 3.1 Flash Image, is Google's fast image generation and editing model for multimodal creative workflows.","icon":"Gemini.Color","tags":"Image,Vision,Reasoning","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"go/gpt-6-astra","description":"GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 3 + c * 15 + cr * 0.3 + cc * 3.75) : tier(\"long_context\", p * 6 + c * 22.5 + cr * 0.6 + cc * 7.5)"},{"model_name":"selfdeploy/clef","quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["Steve3184","free"],"supported_endpoint_types":["systemone"]},{"model_name":"gemini-3.5-flash","description":"Gemini 3.5 Flash is Google's high-efficiency multimodal model, bringing near-Pro level coding and reasoning at Flash-tier cost and speed.","tags":"Vision,Video,Audio,Files,Reasoning,Tools,1M","vendor_id":8,"quota_type":0,"model_ratio":0.75,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["GeminiUltra"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":12,"quota_type":0,"model_ratio":0.045,"model_price":0,"owner_by":"","completion_ratio":3.35,"cache_ratio":0.23,"create_cache_ratio":0.000007,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4-flash","description":"DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.","tags":"Reasoning,Tools,Open Weights,1M","vendor_id":7,"quota_type":0,"model_ratio":0.045,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5\n  \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12)\n    || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.18 + c * 0.72 + cr * 0.0036)\n  : tier(\"off_peak\", p * 0.09 + c * 0.36 + cr * 0.0018)"},{"model_name":"nvidia/llama-3.1-nemoguard-8b-content-safety","description":"Llama 3.1 NemoGuard 8B Content Safety evaluates user prompts and assistant responses against a content-safety policy. It returns safety classifications and categories for moderation workflows.","tags":"Safety","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"claude-sonnet-5","description":"Claude Sonnet 5 is an Anthropic model for coding, agents, and professional work, with adaptive thinking and selectable reasoning effort.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":1.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.6,"enable_groups":["ClaudeR2","ClaudeR3","ClaudeKiro","ClaudeMax","ClaudeR1","ClaudeR4"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 2 + c * 10 + cr * 0.2 + cc * 2.5 + cc1h * 4)"},{"model_name":"oc/nemotron-3.5-super-vl","description":"NVIDIA Nemotron 3.5 Super VL is an open-weight vision-language model with 120B total and 12B active parameters. It supports reasoning, coding, tool use, document understanding, and video understanding, with text, image, and video inputs and text output.","icon":"Nvidia.Color","tags":"Reasoning,Tools,Vision,Video,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/glm-5.2","description":"GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...","icon":"Zhipu.Color","tags":"Reasoning,Tools","vendor_id":12,"quota_type":0,"model_ratio":0.64,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":0.14740348756625,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/kimi-k3.1-pro","description":"Developed by Moonshot AI, Kimi K3.1 Pro succeeds Kimi K3 through additional post-training. This open-weight model combines multimodal understanding with reasoning, accepts text, images and video, and supports tool use within a context window of 1M tokens.","tags":"Vision,Video,Reasoning,Tools,Open Weights,1M","vendor_id":9,"quota_type":0,"model_ratio":1.04,"model_price":0,"owner_by":"","completion_ratio":5.230769230769,"cache_ratio":0.1,"create_cache_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"oc/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":12,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"bge-reranker-v2-m3","description":"BGE Reranker v2 M3 scores the relevance of candidate documents to a query. It reorders retrieved results to improve search and retrieval-augmented generation workflows.","tags":"Rerank","vendor_id":32,"quota_type":0,"model_ratio":0.025,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-2.5-flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,1M","vendor_id":8,"quota_type":0,"model_ratio":0.15,"model_price":0,"owner_by":"","completion_ratio":8.333333333333,"cache_ratio":0.1,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"nvidia/llama-3.1-nemoguard-8b-topic-control","description":"Llama 3.1 NemoGuard 8B Topic Control checks whether user messages stay within the topics and dialogue policies defined for an assistant. It can be used as a topical guardrail in task-focused applications.","tags":"Safety","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"minimax-m3.1-flash","description":"MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":6,"quota_type":0,"model_ratio":0.045,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-opus-4-6","description":"Claude Opus 4.6 is an Anthropic model for coding and long-running professional tasks, designed for agents operating across multi-step workflows.","icon":"Claude.Color","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.6,"enable_groups":["ClaudeKiro","ClaudeMax","ClaudeR1","ClaudeR4","ClaudeR2","ClaudeR3"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"go/minimax-m2.7","description":"MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement. Built to actively participate in its own evolution, M2.7 integrates advanced agentic capabilities through multi-agent...","icon":"Minimax.Color","tags":"Reasoning,Tools","vendor_id":6,"quota_type":0,"model_ratio":0.21,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/mimo-v2.6-pro","description":"MiMo-V2.6-Pro is Xiaomi's foundation model built at a scale of over one trillion parameters.","tags":"Reasoning,Tools","vendor_id":16,"quota_type":0,"model_ratio":0.24,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.008333333333333333,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/kimi-k2.6","description":"Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end coding tasks across Python, Rust, and Go, and...","icon":"Moonshot","tags":"Vision,Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":0.65,"model_price":0,"owner_by":"","completion_ratio":4.153846153846,"cache_ratio":0.2,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/ali/glm-5.3-flashx","description":"GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...","icon":"Zhipu.Color","tags":"Reasoning,Tools,Vision,Video,1M","vendor_id":12,"quota_type":0,"model_ratio":0.45,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.0285,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/gpt-5.6-luna","description":"GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":0.03,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 0.06 + c * 0.36 + cr * 0.006 + cc * 0.075) : tier(\"long_context\", p * 0.12 + c * 0.54 + cr * 0.012 + cc * 0.15)"},{"model_name":"MiniMax-M2.7-free","description":"MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.","tags":"Reasoning,Tools","vendor_id":6,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free"],"supported_endpoint_types":["openai"]},{"model_name":"selfdeploy/gemma-4-31b-it","description":"Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output.","tags":"Vision,Video,Reasoning,Tools,262K","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"selfdeploy/mimo-v2.6-flash","description":"MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi.","tags":"Vision,Video,Audio,Reasoning,Tools,1M","vendor_id":16,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"agnes-video-2.5-flash","description":"Agnes Video 2.5 Flash is a video-generation endpoint for prompt-based visual sequences. It can be used for creative clips, motion concepts, and video prototyping.","tags":"Video","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai-video"]},{"model_name":"meta/llama-3.2-90b-vision-instruct","description":"Llama 3.2 90B Vision Instruct is a multimodal model for understanding images together with text instructions. It can be used for visual question answering, image descriptions, and document or chart interpretation.","tags":"Vision,Files","vendor_id":88,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"glm-5.3-flash-free","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Vision,Reasoning,Tools","vendor_id":12,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free"],"supported_endpoint_types":["openai"]},{"model_name":"codex-auto-review","description":"A code review model intended for use by Codex. It reviews code changes and provides code review feedback.","icon":"Codex.Color","tags":"Reasoning,Tools,Files","vendor_id":2,"quota_type":0,"model_ratio":0.1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"]},{"model_name":"m/gpt-image-2.5-sunburst","description":"GPT Image 2.5 Sunburst is an image generation and editing model for precision-oriented visual work, controlled edits, detailed compositions, and production assets.","icon":"OpenAI","tags":"Image,Vision,Reasoning,Tools","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"step-5-preview","quota_type":0,"model_ratio":0.315,"model_price":0,"owner_by":"","completion_ratio":2.857142857143,"cache_ratio":0.1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/glm-5.1","description":"GLM-5.1 delivers a major leap in coding capability, with particularly significant gains in handling long-horizon tasks. Unlike previous models built around minute-level interactions, GLM-5.1 can work independently and continuously on...","icon":"Zhipu.Color","tags":"Reasoning,Tools","vendor_id":12,"quota_type":0,"model_ratio":0.48,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.21666666666666667,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c 32768 ? tier(\"short\", p * 0.96 + c * 3.84 + cr * 0.208 + cc * 0.00 + cc1h * 0.00) : tier(\"long\", p * 1.28 + c * 4.48 + cr * 0.32 + cc * 0.00 + cc1h * 0.00)"},{"model_name":"oc/beam","quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/kimi-k3","description":"Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI.","tags":"Vision,Open Weights,Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":1.6,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"poolside/laguna-xs-2.1","description":"Laguna XS 2.1 is Poolside's coding agent model in the 33B-A3B category, succeeding Laguna XS.2.","tags":"Reasoning,Tools,Open Weights,262.1K","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/nemotron-3.5-lightning-30b-a3b","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total.","tags":"Reasoning,Tools","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning","description":"NVIDIA Nemotron™ 3 Nano Omni is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems.","tags":"Vision,Reasoning","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"go/qwen3.8-flash","description":"Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.","tags":"Vision,Files,Reasoning,Tools","vendor_id":11,"quota_type":0,"model_ratio":0.08,"model_price":0,"owner_by":"","completion_ratio":3.375,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"hunyuan-mt-7b","description":"Hunyuan MT 7B is a machine-translation model for converting text between languages. It can be used for multilingual documents, localization, and cross-language communication.","tags":"8.2K","vendor_id":69,"quota_type":0,"model_ratio":0.0015,"model_price":0,"owner_by":"","completion_ratio":3.3333333333333335,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"kolors","description":"Kolors is a text-to-image model for generating visual content from natural-language prompts. It can be used for illustrations, concept art, and creative image generation.","tags":"Image","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"go/mimo-v2.6-flash","description":"MiMo-V2.6-Flash is Xiaomi's open-source Mixture-of-Experts foundation model, with 309B total parameters and 15B activated per token.","tags":"Open Weights,Reasoning,Tools","vendor_id":16,"quota_type":0,"model_ratio":0.2,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.03,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"qwen3.7-max","description":"Qwen3.7-Max is a text model in Alibaba's Qwen3.7 series, designed for agent-centric workloads including coding, office work, and productivity tasks.","tags":"Reasoning,Tools,1M","vendor_id":11,"quota_type":0,"model_ratio":0.54,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/nemotron-3.5-content-safety","description":"NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B.","tags":"Safety,Vision","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"ali/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...","icon":"Zhipu.Color","tags":"Reasoning,Tools,Vision,Video,1M","vendor_id":12,"quota_type":0,"model_ratio":0.1,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.2875,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"m/gpt-image-2.5-flare","description":"GPT Image 2.5 Flare is an image generation and editing model designed for rapid visual iteration, creative exploration, and repeated image editing workflows.","icon":"OpenAI","tags":"Image,Vision,Reasoning,Tools","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"claude-opus-4-8","description":"Claude Opus 4.8 is an Anthropic reasoning model in the Opus family, supporting text, image, and file inputs with text output.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.6,"enable_groups":["ClaudeR2","ClaudeR3","ClaudeKiro","ClaudeMax","ClaudeR1","ClaudeR4"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"go/gpt-5.6-sol","description":"GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":0.75,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 1.5 + c * 9 + cr * 0.15 + cc * 1.875) : tier(\"long_context\", p * 3 + c * 13.5 + cr * 0.3 + cc * 3.75)"},{"model_name":"glm-5.2","description":"GLM 5.2 is a large-scale reasoning model from Z.ai.","tags":"Reasoning,Tools,Open Weights,1M","vendor_id":12,"quota_type":0,"model_ratio":0.36,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"kimi-k3","description":"Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI.","tags":"Vision,Open Weights,Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":0.9,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-opus-4-7","description":"Opus 4.7 is the next generation of Anthropic's Opus family, built for long-running, asynchronous agents.","icon":"Claude.Color","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.6,"enable_groups":["ClaudeMax","ClaudeR1","ClaudeR4","ClaudeR2","ClaudeR3","ClaudeKiro"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"claude-opus-5-5","description":"Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code review and bug finding, financial and scientific analysis, and reading dense charts, diagrams, and screenshots, and it is more careful than its predecessor about only stating figures and citing sources it can back up. The model completes comparable tasks in fewer steps and with fewer tokens than Opus 5, and reports on its work in plainer language, with clear updates on what it did, what it found, and what it needs from the user. Thinking is always adaptive, so effort is the main lever for trading off depth, latency, and cost, and lower effort settings remain effective for latency-sensitive workloads.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["ClaudeR1","ClaudeR4","ClaudeR2","ClaudeAnti","ClaudeR3","ClaudeKiro","ClaudeMax"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 4 + c * 20 + cr * 0.2 + cc1h * 8 + cc * 5)"},{"model_name":"selfdeploy/omnivoice","quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"go/ali/glm-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...","icon":"Zhipu.Color","tags":"Reasoning,Tools,1M","vendor_id":12,"quota_type":0,"model_ratio":1.8,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"bge-large-zh-v1.5","description":"BGE Large ZH v1.5 converts Chinese text into vector embeddings for semantic search and similarity matching. It can be used for document retrieval, clustering, and retrieval-augmented generation.","tags":"Embeddings","vendor_id":32,"quota_type":0,"model_ratio":0.01,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"google/diffusiongemma-26b-a4b-it","description":"DiffusionGemma 26B A4B IT is an instruction-tuned Gemma model that generates text using diffusion. It can be used for conversational assistance, drafting, and text-completion workflows.","icon":"Gemma.Color","tags":"Reasoning,Open Weights","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"m/nano-banana-2-lite","description":"Nano Banana 2 Lite is the upstream Lite variant of the Nano Banana 2 image-generation service. It creates images from prompts and reference images; available output settings follow this service route.","icon":"Gemini.Color","tags":"Image,Vision","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"kimi-k3.1-pro","description":"Developed by Moonshot AI, Kimi K3.1 Pro succeeds Kimi K3 through additional post-training. This open-weight model combines multimodal understanding with reasoning, accepts text, images and video, and supports tool use within a context window of 1M tokens.","tags":"Vision,Video,Reasoning,Tools,Open Weights,1M","vendor_id":9,"quota_type":0,"model_ratio":0.585,"model_price":0,"owner_by":"","completion_ratio":5.230769230769,"cache_ratio":0.1,"create_cache_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"go/grok-4.7","description":"Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...","icon":"XAI","tags":"Vision,Files,Reasoning,Tools","vendor_id":5,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 0.16 + c * 0.48 + cr * 0.04) : tier(\"long_context\", p * 0.32 + c * 0.96 + cr * 0.08)"},{"model_name":"selfdeploy/d3-flash","quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["Steve3184","free"],"supported_endpoint_types":["systemone"]},{"model_name":"go/minimax-m3.1-flash","description":"MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":6,"quota_type":0,"model_ratio":0.08,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"m/gemini-flash","description":"Gemini Flash Latest is an alias for the latest Google Gemini Flash model provided by the upstream service. It supports multimodal understanding, reasoning, and tool use; the concrete model version follows the upstream mapping.","icon":"Gemini.Color","tags":"Vision,Files,Audio,Video,Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.004,"owner_by":"","completion_ratio":0,"enable_groups":["Gemini"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.1-pro-preview","description":"Gemini 3.1 Pro Preview is Google’s frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows.","tags":"Reasoning,Tools","vendor_id":8,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["GeminiUltra"],"supported_endpoint_types":["gemini","openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4)"},{"model_name":"hy4-preview","description":"Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total.","tags":"Reasoning,Tools,Open Weights","vendor_id":69,"quota_type":0,"model_ratio":0.27,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.05,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4-flash-0731","description":"DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.","tags":"Vision,Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":0.045,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5\n  \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12)\n    || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.18 + c * 0.72 + cr * 0.0036)\n  : tier(\"off_peak\", p * 0.09 + c * 0.36 + cr * 0.0018)"},{"model_name":"mimo-v2.6-pro","description":"MiMo-V2.6-Pro is Xiaomi's foundation model built at a scale of over one trillion parameters.","tags":"Reasoning,Tools","vendor_id":16,"quota_type":0,"model_ratio":0.135,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.008333333333333333,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-4-argon","description":"Gemini 4 Argon is a Google multimodal reasoning model with a unified 1M-token context window. It accepts text, images, video, audio, and PDF inputs and generates text, with support for function calling, structured output, and adjustable thinking levels.","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,Video,1M","vendor_id":8,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.05,"enable_groups":["GeminiUltra"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"selfdeploy/qwen3.8-flash-next","vendor_id":-1022,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"selfdeploy/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":12,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"veo-3.1-lite","quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", u(\"video_count\") * 0.2)","billing_usage_schema":{"video_count":{"type":"number","unit":"count","unitLabel":{"en":"video","zh":"条"},"description":{"en":"Video generation unit price","zh":"视频生成单价"}}},"billing_usage_examples":[{"label":"1 video","facts":{"video_count":1}}]},{"model_name":"gemini-3.6-flash","description":"Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development.","tags":"Vision,Video,Audio,Files,Reasoning,Tools,1M","vendor_id":8,"quota_type":0,"model_ratio":0.75,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["GeminiUltra"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"nvidia/llama-3.1-nemotron-safety-guard-8b-v3","description":"Llama 3.1 Nemotron Safety Guard 8B v3 classifies user prompts and model responses for content-safety review. It identifies unsafe content and reports relevant policy categories.","tags":"Safety","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/riva-translate-4b-instruct-v2","description":"Riva Translate 4B Instruct v2 is a multilingual text-translation model with instruction-based language selection. It can be used for document translation, localization, and cross-language text processing.","tags":"Translation","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-5.6-sol","description":"GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":2,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["CodexCyber","CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 5 + c * 30 + cr * 0.5 + cc * 6.25) : tier(\"long_context\", p * 10 + c * 45 + cr * 1 + cc * 12.5)"},{"model_name":"go/hy4-preview","description":"Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...","icon":"Hunyuan.Color","tags":"Reasoning,Tools","vendor_id":69,"quota_type":0,"model_ratio":0.48,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.05,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/kimi-k2.8-preview","vendor_id":9,"quota_type":0,"model_ratio":0.52,"model_price":0,"owner_by":"","completion_ratio":4.153846153846154,"cache_ratio":0.26153846153846155,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"grok-4.7-build-fast","description":"Grok 4.7 Build Fast is a coding-oriented model endpoint for software-development assistance. It can be used for code generation, debugging, code explanation, and implementation tasks.","tags":"Reasoning,Tools","vendor_id":5,"quota_type":0,"model_ratio":2,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["Grok"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 4 + c * 12 + cr * 1) : tier(\"long_context\", p * 6 + c * 18 + cr * 1.5)"},{"model_name":"kling-o3","vendor_id":-1017,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", u(\"video_count\") * 0.65)","billing_usage_schema":{"video_count":{"type":"number","unit":"count","unitLabel":{"en":"video","zh":"条"},"description":{"en":"Video generation unit price","zh":"视频生成单价"}}},"billing_usage_examples":[{"label":"1 video","facts":{"video_count":1}}]},{"model_name":"go/longcat-2.5-preview-free","description":"LongCat 2.5 Preview is a preview endpoint in the LongCat family for conversational assistance, text generation, and coding tasks. Features and behavior may change during the preview.","tags":"Reasoning,Tools","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"mimo-v2.5","description":"MiMo-V2.5 is Xiaomi's omnimodal model for agentic tasks and multimodal perception, including image and video understanding.","icon":"XiaomiMiMo","tags":"Reasoning,Tools,Files,Open Weights,Vision,Audio,1M","vendor_id":16,"quota_type":0,"model_ratio":0.075,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.1,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"ali/kimi-k3","description":"Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI.","icon":"Moonshot","tags":"Vision,Open Weights,Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":0.075,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"oc/deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture.","icon":"DeepSeek.Color","tags":"Reasoning,Tools","vendor_id":7,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"grok-4.7","description":"Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6.","tags":"Reasoning,Tools","vendor_id":5,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Grok","free"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 1.6 + c * 4.8 + cr * 0.4) : tier(\"long_context\", p * 3.2 + c * 9.6 + cr * 0.8)"},{"model_name":"gpt-image-2","description":"GPT Image 2 is an image generation and editing model that works with text prompts and reference images. It can be used for illustrations, visual design, and image transformation workflows.","icon":"OpenAI","tags":"Files,Vision","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.055,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"gpt-image-2.5-flare","description":"GPT Image 2.5 Flare is an image generation and editing model oriented toward rapid visual iteration. It can be used for image drafts, creative exploration, and repeated editing workflows.","tags":"Image,Reasoning,Tools","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.055,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"go/muse-spark-1.3-contributor","description":"Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows.","tags":"Vision,Reasoning,Tools","vendor_id":88,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"go/space-bunny-free","description":"Space Bunny Free is an anonymous OpenCode free model for lightweight coding and agent workflows. It is offered as a limited free endpoint and may change over time.","tags":"Reasoning,Tools","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"ali/glm-5.3-flashx","description":"GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...","icon":"Zhipu.Color","tags":"Reasoning,Tools,Vision,Video,1M","vendor_id":12,"quota_type":0,"model_ratio":0.25,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.0285,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-6.1-sol","description":"GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":2,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 10 + cr * 0.1 + cc * 2.5) : tier(\"long_context\", p * 4 + c * 15 + cr * 0.2 + cc * 5)"},{"model_name":"cohere-north-mini-code-free","description":"North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized for code generation, agentic software engineering, and terminal tasks, and is trained to generalize across agent harnesses such as OpenCode and SWE-Agent. It offers a 256K-token context window with up to 64K tokens of output, supports interleaved reasoning and tool use via JSON schema, and is released open-weight under the Apache 2.0 license. Its small active-parameter footprint enables low-latency inference, including on local hardware.","icon":"Cohere.Color","tags":"Reasoning,Tools,Open Weights,256K","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free"],"supported_endpoint_types":["openai"]},{"model_name":"go/qwen3.8-max","description":"Qwen3.8 Max is a general-purpose Qwen model for reasoning, coding, document analysis, and written content. The exact model snapshot follows the configured upstream service.","tags":"Files,Reasoning,Tools","vendor_id":11,"quota_type":0,"model_ratio":0.96,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"m/nano-banana-pro","description":"Nano Banana Pro is Google's high-quality image generation and editing model for detailed visual creation, compositing, and controlled image transformation.","icon":"Gemini.Color","tags":"Image,Vision,Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0005,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"bge-large-en-v1.5","description":"BGE Large EN v1.5 converts English text into vector embeddings for semantic search and similarity matching. It can be used for document retrieval, clustering, and retrieval-augmented generation.","tags":"Embeddings","vendor_id":32,"quota_type":0,"model_ratio":0.01,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"openai/gpt-oss-20b","description":"gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.","tags":"Open Weights,Reasoning,Tools","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"oc/cat-jam","quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-fable-5","description":"Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":2,"enable_groups":["ClaudeR1","ClaudeR4","ClaudeR2","ClaudeR3"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 10 + c * 50 + cr * 1 + cc * 12.5 + cc1h * 20)"},{"model_name":"grok-imagine-image","description":"Grok Imagine Image is an image-generation endpoint in the Grok family for creating visual content from text prompts. It can be used for illustrations, concept images, and creative assets.","tags":"Files,Vision,8K","vendor_id":5,"quota_type":1,"model_ratio":0,"model_price":0.05,"owner_by":"","completion_ratio":0,"enable_groups":["free","Grok"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-2.5-flash-lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,1M","vendor_id":8,"quota_type":0,"model_ratio":0.05,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"go/claude-haiku-5-5","description":"Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and knowledge work, and accepts text and image input with a 1M-token context window. It is the first Haiku model with adjustable effort. Thinking is adaptive and on by default, with effort as the main lever for trading off depth, latency, and cost; it can be turned off at low, medium, and high effort.","icon":"Claude.Color","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 100000 ? tier(\"standard\", p * 0.1000000 + c * 0.5000000 + cr * 0.01000000 + cc * 0.125000000 + cc1h * 0.2000000) : tier(\"long_context\", p * 0.5000000 + c * 2.5000000 + cr * 0.05000000 + cc * 0.625000000 + cc1h * 1.000000)"},{"model_name":"jev-1.13","description":"Jev 1.13 is a structured decision model from TypeSafe for software workflows that need typed choices or constrained decisions rather than open-ended prose.","tags":"Reasoning","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0.00002,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["systemone"]},{"model_name":"go/minimax-m3","description":"MiniMax-M3 is a multimodal foundation model from MiniMax that accepts text, image, and video inputs and produces text, with support for long-horizon agentic work and coding.","tags":"Vision,Reasoning,Tools","vendor_id":6,"quota_type":0,"model_ratio":0.168,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"claude-sonnet-4-5-20250929","description":"Claude Sonnet 4.5 is an Anthropic model optimized for agentic applications and coding workflows.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":1,"quota_type":1,"model_ratio":0,"model_price":0.1,"owner_by":"","completion_ratio":0,"enable_groups":["ClaudeMax","ClaudeR1","ClaudeKiro (Sonnet 4.5)"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 3 + c * 15 + cr * 0.3 + cc * 3.75 + cc1h * 6)"},{"model_name":"telespeechasr","description":"TeleSpeechASR is an automatic speech-recognition model for converting spoken audio into text. It can be used for transcription, searchable audio records, and downstream text processing.","tags":"Audio","vendor_id":99,"quota_type":0,"model_ratio":0.000075,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-opus-4-5-20251101","description":"Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":5,"enable_groups":["ClaudeKiro"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5 \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.18 + c * 0.72 + cr * 0.0036)\n  : tier(\"off_peak\", p * 0.09 + c * 0.36 + cr * 0.0018)"},{"model_name":"ali/glm-5.2","description":"GLM 5.2 is a large-scale reasoning model from Z.ai.","icon":"Zhipu.Color","tags":"Reasoning,Tools,Open Weights,1M","vendor_id":12,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-sonnet-5-5","description":"Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing polished documents, slides, and spreadsheets, and it writes and communicates more clearly than its predecessor. Thinking is always on, so effort is the main lever for trading off depth, latency, and cost, and lower effort settings keep it responsive for everyday agentic loops.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["ClaudeMax","ClaudeR1","ClaudeR4","ClaudeR2","ClaudeAnti","ClaudeR3"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 2.000000 + c * 10.00000 + cr * 0.1000000 + cc * 2.5000000 + cc1h * 4.000000)"},{"model_name":"grok-4.6","description":"Grok 4.6 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM. It is succeeded by Grok 4.7.","tags":"Reasoning,Tools,Files,Vision,500K","vendor_id":5,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["Grok"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 6 + cr * 0.5) : tier(\"long_context\", p * 4 + c * 12 + cr * 1)"},{"model_name":"qwen3-8b","description":"Qwen3-8B is a dense 8.2B parameter causal language model from the Qwen3 series, designed for both reasoning-heavy tasks and efficient dialogue.","tags":"Reasoning,Tools,131K","vendor_id":11,"quota_type":0,"model_ratio":0.002,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"billing_plugin_variants":[{"plugin_key":"comfyui-1","plugin_name":"ComfyUI","billing_expr":"u(\"resolution\") == \"480p\" ? tier(\"480p\", u(\"seconds\") * 0.055) : u(\"resolution\") == \"768p\" ? tier(\"768p\", u(\"seconds\") * 0.075) : u(\"resolution\") == \"1088p\" ? tier(\"1088p\", u(\"seconds\") * 0.085) : tier(\"1440p\", u(\"seconds\") * 0.095)","billing_mode":"tiered_expr","billing_usage_schema":{"resolution":{"enum":["480p","768p","1088p","1440p"],"description":{"en":"Output video resolution","zh":"输出视频分辨率"}},"seconds":{"type":"number","unit":"second","description":{"en":"Video generation unit price","zh":"视频生成单价"}}},"billing_usage_examples":[{"label":"480p · 5s","facts":{"resolution":"480p","seconds":5}},{"label":"768p · 5s","facts":{"resolution":"768p","seconds":5}},{"label":"1088p · 5s","facts":{"resolution":"1088p","seconds":5}},{"label":"1440p · 5s","facts":{"resolution":"1440p","seconds":5}}]}],"model_name":"MiniMax-H3","description":"MiniMax H3 is a video-generation model from MiniMax for creating short videos from text prompts and reference images.","icon":"Minimax.Color","tags":"Video,Image","vendor_id":6,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"u(\"resolution\") == \"480p\" ? tier(\"480p\", u(\"seconds\") * 0.055) : u(\"resolution\") == \"768p\" ? tier(\"768p\", u(\"seconds\") * 0.075) : u(\"resolution\") == \"1088p\" ? tier(\"1088p\", u(\"seconds\") * 0.085) : tier(\"1440p\", u(\"seconds\") * 0.095)","billing_usage_schema":{"resolution":{"enum":["480p","768p","1088p","1440p"],"description":{"en":"Output video resolution","zh":"输出视频分辨率"}},"seconds":{"type":"number","unit":"second","description":{"en":"Video generation unit price","zh":"视频生成单价"}}},"billing_usage_examples":[{"label":"480p · 5s","facts":{"resolution":"480p","seconds":5}},{"label":"768p · 5s","facts":{"resolution":"768p","seconds":5}},{"label":"1088p · 5s","facts":{"resolution":"1088p","seconds":5}},{"label":"1440p · 5s","facts":{"resolution":"1440p","seconds":5}}]},{"model_name":"azure-tts","description":"Azure TTS is a speech-synthesis endpoint for converting text into spoken audio. Available voices, languages, and output formats depend on the configured upstream service.","tags":"Audio","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"ali/glm-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...","icon":"Zhipu.Color","tags":"Reasoning,Tools,1M","vendor_id":12,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"oc/mimo-v2.6-flash","description":"MiMo-V2.6-Flash is Xiaomi's open-source Mixture-of-Experts foundation model, with 309B total parameters and 15B activated per token.","tags":"Open Weights,Reasoning,Tools","vendor_id":16,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"go/kimi-k2.7-code","description":"MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts.","tags":"Reasoning,Tools","vendor_id":9,"quota_type":0,"model_ratio":0.65,"model_price":0,"owner_by":"","completion_ratio":4.153846153846,"cache_ratio":0.2,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"meta/llama-guard-4-12b","description":"Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification.","tags":"Safety,Vision","vendor_id":88,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"nvidia/nemotron-3-ultra-550b-a55b","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE).","tags":"Reasoning,Tools,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"z-ai/glm-5.3-flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks.","tags":"Vision,Reasoning,Tools","vendor_id":12,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"grok-4.5","description":"Grok 4.5 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM.","tags":"Reasoning,Tools,Files,Vision,500K","vendor_id":5,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["Grok"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 6 + cr * 0.3) : tier(\"long_context\", p * 4 + c * 12 + cr * 0.6)"},{"model_name":"glm-4-9b-0414","description":"GLM 4 9B 0414 is an instruction-following language model for conversational assistance and text tasks. It can be used for question answering, drafting, summarization, and coding support.","tags":"32K","vendor_id":12,"quota_type":0,"model_ratio":0.0015,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"enable_groups":["free","default"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-ai/deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"agnes-image-2.1-flash","description":"Agnes Image 2.1 Flash is an image-generation endpoint for creating visual content from prompts. It can be used for illustrations, design concepts, and creative assets.","tags":"Image","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.000004,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai-video"]},{"model_name":"agnes-3.0-flash","description":"Agnes 3.0 Flash is a text-generation endpoint for conversational assistance and written content. It can be used for question answering, drafting, summarization, and text transformation.","tags":"Text","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.000004,"owner_by":"","completion_ratio":0,"enable_groups":["free","default"],"supported_endpoint_types":["openai-video"]},{"model_name":"glm-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.","tags":"Reasoning,Tools,1M","vendor_id":12,"quota_type":0,"model_ratio":0.36,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"google/gemma-4-31b-it","description":"Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output.","icon":"Gemma.Color","tags":"Reasoning,Tools,Files,Vision,262.1K","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["Steve3184","NIM"],"supported_endpoint_types":["openai"]},{"model_name":"meta/llama-3.2-11b-vision-instruct","description":"Llama 3.2 11B Vision Instruct is a multimodal model that combines image understanding with text instructions. It can be used for image descriptions, visual question answering, and image-based assistance.","tags":"Vision","vendor_id":88,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"glm-5.1","description":"GLM-5.1 delivers a major leap in coding capability, with particularly significant gains in handling long-horizon tasks.","tags":"Reasoning,Tools,Open Weights,200K","vendor_id":12,"quota_type":0,"model_ratio":0.36,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-fable-5-1","description":"Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual code generation, and finance and analysis tasks in particular. It also tends to be more concise than Fable 5 in its plans and summaries. We recommend testing it as a direct upgrade wherever you use Fable 5 today, and alongside Opus 5 on reasoning-heavy tasks.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["ClaudeR3","ClaudeR1","ClaudeR4","ClaudeR2"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 10.00000 + c * 50.00000 + cr * 0.25000000 + cc * 12.5000000 + cc1h * 20.00000)"},{"model_name":"nvidia/nemotron-parse-2.0","description":"NVIDIA Nemotron Parse 2.0 converts document images into structured text, layout classes, bounding boxes, and reading order. It supports multilingual OCR, handwritten-text extraction, table recovery, and chart-to-table parsing for document understanding and information extraction.","icon":"Nvidia.Color","tags":"Vision,Files,Open Weights","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["Steve3184","NIM"],"supported_endpoint_types":["openai"]},{"model_name":"agnes-image-2.5-flash","description":"Agnes Image 2.5 Flash is an image-generation endpoint for creating visual content from prompts. It can be used for illustrations, design concepts, and creative assets.","tags":"Image","vendor_id":98,"quota_type":1,"model_ratio":0,"model_price":0.000004,"owner_by":"","completion_ratio":0,"enable_groups":["default","free"],"supported_endpoint_types":["openai-video"]},{"model_name":"claude-opus-5","description":"Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["ClaudeR3","ClaudeKiro","ClaudeMax","ClaudeR1","ClaudeR4","ClaudeR2"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"selfdeploy/deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture.","tags":"Vision,Reasoning,Tools,1M","vendor_id":7,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"go/mimo-v2.5-pro","description":"MiMo-V2.5-Pro is a Xiaomi model for agentic applications, complex software engineering, and long-horizon tasks.","tags":"Reasoning,Tools","vendor_id":16,"quota_type":0,"model_ratio":0.3,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.008333333333333333,"create_cache_ratio":0,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-ocr","description":"DeepSeek OCR is a document-reading model for extracting text and structured content from images. It can be used for document digitization, information extraction, and preparing material for search or analysis.","tags":"Files,Open Weights,Vision,8.2K","vendor_id":7,"quota_type":0,"model_ratio":0.0025,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"anti/gemini-3.8-flash","description":"Gemini 3.8 Flash is a Google model for software engineering, agentic tasks, and multi-step reasoning, building on Gemini 3.7 Flash.","tags":"Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.0375,"owner_by":"","completion_ratio":0,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"go/gpt-6-luna","description":"GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic tasks, and at higher reasoning effort it can take on complex software engineering and computer-use tasks that previously called for a Sol-tier model. It shares the GPT-6 family's gains in factual reliability and its clearer, more concise communication style.","tags":"Vision,Files,Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 0.034 + cr * 0.0034 + cc * 0.04250 + c * 0.170) : tier(\"long_context\", p * 0.068 + cr * 0.0068 + cc * 0.0850 + c * 0.2550)"},{"model_name":"MiniMax-M2.7-highspeed","description":"MiniMax-M2.7-highspeed is the faster-inference version of MiniMax-M2.7, retaining the same model capabilities. It is designed for software engineering, multilingual code generation, code refactoring, reasoning, and tool-based agent workflows.","icon":"Minimax.Color","tags":"Reasoning,Tools","vendor_id":6,"quota_type":0,"model_ratio":0.315,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.1,"create_cache_ratio":0.625,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"claude-haiku-5-5","description":"Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and knowledge work, and accepts text and image input with a 1M-token context window. It is the first Haiku model with adjustable effort. Thinking is adaptive and on by default, with effort as the main lever for trading off depth, latency, and cost; it can be turned off at low, medium, and high effort.","icon":"Claude.Color","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":1,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["ClaudeR1","ClaudeR4","ClaudeR2","ClaudeR3","ClaudeMax"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 100000 ? tier(\"standard\", p * 0.1000000 + c * 0.5000000 + cr * 0.01000000 + cc * 0.125000000 + cc1h * 0.2000000) : tier(\"long_context\", p * 0.5000000 + c * 2.5000000 + cr * 0.05000000 + cc * 0.625000000 + cc1h * 1.000000)"},{"model_name":"selfdeploy/d3","quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["systemone"]},{"model_name":"kling-v3","vendor_id":-1017,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", u(\"video_count\") * 0.65)","billing_usage_schema":{"video_count":{"type":"number","unit":"count","unitLabel":{"en":"video","zh":"条"},"description":{"en":"Video generation unit price","zh":"视频生成单价"}}},"billing_usage_examples":[{"label":"1 video","facts":{"video_count":1}}]},{"model_name":"go/mimo-v2.5","description":"MiMo-V2.5 is Xiaomi's omnimodal model for agentic tasks and multimodal perception, including image and video understanding.","tags":"Vision,Reasoning,Tools","vendor_id":16,"quota_type":0,"model_ratio":0.1,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.1,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"]},{"model_name":"meta/muse-glimmer-30b","description":"Muse Glimmer 30B is Meta's open-weight multimodal model for autonomous agents. It combines text and image understanding with multi-step reasoning, tool use, coding, and recovery from tool failures, and is designed for local deployment on consumer hardware.","icon":"Meta.Color","tags":"Reasoning,Tools,Vision,Open Weights","vendor_id":88,"quota_type":1,"model_ratio":0,"model_price":0.0001,"owner_by":"","completion_ratio":0,"enable_groups":["NIM","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4-pro","description":"DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window.","tags":"Reasoning,Tools,Open Weights,1M","vendor_id":7,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5 \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.81 + c * 2.43 + cr * 0.027)\n  : tier(\"off_peak\", p * 0.405 + c * 1.215 + cr * 0.0135)"},{"model_name":"go/deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5 \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)))\n  ? tier(\"peak\", p * 0.32 + c * 1.28 + cr * 0.0064)\n  : tier(\"off_peak\", p * 0.16 + c * 0.64 + cr * 0.0032)"},{"model_name":"oc/nemotron-3.5-lightning","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total.","tags":"Reasoning,Tools","vendor_id":92,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"oc/big-pickle","description":"Big Pickle is a model endpoint exposed through OpenCode for coding assistance and development tasks. The public alias does not identify a specific underlying base-model version.","icon":"OpenCode","tags":"Reasoning,Tools","vendor_id":99,"quota_type":1,"model_ratio":0,"model_price":0.00005,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"grok-imagine-image-2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","icon":"XAI.Color","tags":"Files,Vision,8K","vendor_id":5,"quota_type":1,"model_ratio":0,"model_price":0.08,"owner_by":"","completion_ratio":0,"enable_groups":["free","Grok"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3-pro-image","description":"Gemini 3 Pro Image, also known as Nano Banana Pro, is Google's image generation and editing model built on Gemini 3 Pro.","tags":"Image,Reasoning,Tools","vendor_id":8,"quota_type":1,"model_ratio":0,"model_price":0.05,"owner_by":"","completion_ratio":0,"enable_groups":["default"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"gemini-3.5-flash-lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","icon":"Gemini.Color","tags":"Reasoning,Tools,Files,Vision,Audio,1M","vendor_id":8,"quota_type":0,"model_ratio":0.15,"model_price":0,"owner_by":"","completion_ratio":8.333333333333334,"cache_ratio":0.1,"enable_groups":["Gemini"],"supported_endpoint_types":["gemini","openai"]},{"model_name":"glm-5.3-flashx","description":"GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model.","tags":"Vision,Reasoning,Tools","vendor_id":12,"quota_type":0,"model_ratio":0.15,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.285,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"grok-build-0.1","description":"Grok Build 0.1 is SpaceXAI’s fast coding model trained specifically for agentic software engineering workflows.","tags":"Reasoning,Tools,Files,Vision,256K","vendor_id":5,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Grok"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 1 + c * 2 + cr * 0.2) : tier(\"long_context\", p * 2 + c * 4 + cr * 0.4)"},{"model_name":"selfdeploy/qwen3.8-27b","description":"Qwen3.8 27B is an open-weight dense vision-language model from Qwen.","tags":"Vision,Video,Reasoning,Tools,1M","vendor_id":11,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free","Steve3184"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-6-sol","description":"GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":2,"enable_groups":["CodexPro","CodexTeam"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 10 + cr * 0.2 + cc * 2.5) : tier(\"long_context\", p * 4 + c * 15 + cr * 0.4 + cc * 5)"},{"model_name":"deepseek-r1-0528-qwen3-8b","description":"DeepSeek R1 0528 Qwen3 8B is a compact reasoning model for multi-step problem solving, coding, and mathematical tasks. It combines the Qwen3 8B model family with reasoning capabilities from DeepSeek R1.","tags":"Reasoning,Tools","vendor_id":7,"quota_type":0,"model_ratio":0.0025,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["default","free"],"supported_endpoint_types":["openai"]},{"model_name":"indextts-2","description":"IndexTTS-2 is an autoregressive text-to-speech model from IndexTeam, with expressive speech synthesis and emotion control through multiple input modalities.","tags":"Audio","vendor_id":99,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", u(\"requests\") * 0.015 + u(\"seconds\") * 0.002)","billing_usage_schema":{"requests":{"type":"number","unit":"count","description":{"en":"Minimum audio generation unit price","zh":"音频生成最低单价"}},"seconds":{"type":"number","unit":"second","description":{"en":"Audio generation unit price","zh":"音频生成单价"}}},"billing_usage_examples":[{"label":"Short audio","facts":{"requests":1,"seconds":1}},{"label":"Longer audio","facts":{"requests":1,"seconds":10}}]},{"model_name":"openai/gpt-5.6-luna","description":"GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":2,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free"],"supported_endpoint_types":["openai"]},{"model_name":"go/gpt-6.1-sol","description":"GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series.","tags":"Reasoning,Tools","vendor_id":2,"quota_type":0,"model_ratio":37.5,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["Go Plan"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 0.64 + c * 3.20 + cr * 0.032 + cc * 0.800) : tier(\"long_context\", p * 1.28 + c * 4.80 + cr * 0.064 + cc * 1.60)"},{"model_name":"deepseek-v4-flash-0731-free","description":"DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.","tags":"Vision,Reasoning,Tools","vendor_id":7,"quota_type":1,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":0,"enable_groups":["free"],"supported_endpoint_types":["openai"]},{"model_name":"qwen3.8-max","description":"Qwen3.8 Max is a general-purpose Qwen model for reasoning, coding, document analysis, and written content. The exact model snapshot follows the configured upstream service.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":11,"quota_type":0,"model_ratio":0.54,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.2,"create_cache_ratio":1.25,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]},{"model_name":"hy3","description":"Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use.","tags":"Reasoning,Tools,Open Weights,256K","vendor_id":69,"quota_type":0,"model_ratio":0.045,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.25,"create_cache_ratio":1,"image_ratio":1,"enable_groups":["default"],"supported_endpoint_types":["openai"]}],"group_ratio":{"ClaudeAnti":0.55,"ClaudeKiro":0.25,"ClaudeKiro (Sonnet 4.5)":0.04,"ClaudeMax":0.85,"ClaudeR1":0.47,"ClaudeR2":0.47,"ClaudeR3":0.47,"ClaudeR4":0.47,"CodexPro":0.2,"CodexTeam":0.175,"CodexTeam2":0.075,"Gemini":0.2,"GeminiUltra":0.4,"Go Plan":1,"Grok":0.12,"NIM":0,"default":1,"free":0},"model_square_group_descriptions":{"Go Plan":"🔵订阅分组 购买后选择该分组来使用Go Plan 模型广场中价格÷2为实际等效额度","free":""},"model_square_groups":["Go Plan","free"],"pricing_version":"a42d372ccf0b5dd13ecf71203521f9d2","success":true,"supported_endpoint":{"anthropic":{"path":"/v1/messages","method":"POST"},"gemini":{"path":"/v1beta/models/{model}:generateContent","method":"POST"},"image-generation":{"path":"/v1/images/generations","method":"POST"},"openai":{"path":"/v1/chat/completions","method":"POST"},"systemone":{"path":"/v1/systemone","method":"POST"}},"usable_group":{"ClaudeAnti":"🟢Antigravitiy (Ultra) 渠道 智商和ccmax一样 有缓存","ClaudeKiro":"🟢正价Pro渠道 保真 稳定性一般 90%虚拟缓存","ClaudeKiro (Sonnet 4.5)":"🟠Free号 只有Sonnet4.5","ClaudeMax":"🟢Max 20x号池 超高缓存(1h) 限制cc客户端 禁止蒸馏 破限\n     禁止生物等任何能触发refusal的使用，触发默认按 3x ~ 5x计费","ClaudeR1":"🟠逆向Claude 渠道1 只有部分模型是纯的 不是kiro/cursor 有缓存","ClaudeR2":"🟠逆向Claude 渠道2 只有部分模型是纯的 不是kiro/cursor 有缓存","ClaudeR3":"🟠逆向Claude 渠道3 只有部分模型是纯的 不是kiro/cursor 有缓存","ClaudeR4":"🟠逆向Claude 渠道4 只有部分模型是纯的 不是kiro/cursor 有缓存","CodexPro":"🟢正价Pro 不降智 稳定性优先 高缓存","CodexTeam":"🟢Team 5x不降智渠道 缓存一般","CodexTeam2":"🟠Team 5x/1x速刷 因为都降智成Luna了故只开这个模型","Gemini":"🟠Antigravitiy (Pro), GeminiCLI 有无缓存自测","GeminiUltra":"🟠Antigravitiy (Ultra) 比较稳定但缓存一般 推荐只用里面的Argon","Grok":"🟢Heavy 渠道 稳定性95%","NIM":"🔴Nvidia NIM 渠道 免费 不限速 可直接买号","default":"🟢包含TTS,生图,视频生成,中国模型等"},"vendors":[{"id":72,"name":"Kolors","icon":"Kolors.Color"},{"id":146,"name":"OpenBMB"},{"id":11,"name":"Alibaba","icon":"Qwen.Color"},{"id":12,"name":"Zhipu","icon":"Zhipu.Color"},{"id":142,"name":"IndexTeam"},{"id":-1022,"name":"阿里巴巴","icon":"Qwen.Color"},{"id":5,"name":"xAI","icon":"XAI"},{"id":98,"name":"AgnesAI","icon":"AgnesAI"},{"id":144,"name":"Meituan"},{"id":2,"name":"OpenAI","icon":"OpenAI"},{"id":16,"name":"Xiaomi","icon":"XiaomiMiMo"},{"id":143,"name":"Stepfun"},{"id":10,"name":"Mistral","icon":"Mistral.Color"},{"id":96,"name":"Poolside","icon":"Poolside.Color"},{"id":-1017,"name":"快手","icon":"Kling.Color"},{"id":1,"name":"Anthropic","icon":"Claude.Color"},{"id":32,"name":"BAAI","icon":"BAAI"},{"id":58,"name":"Azure","icon":"AzureAI.Color"},{"id":69,"name":"Tencent","icon":"Hunyuan.Color"},{"id":140,"name":"Bytedance","icon":"ByteDance.Color"},{"id":141,"name":"InclusionAI"},{"id":145,"name":"Cohere","icon":"Cohere.Color"},{"id":7,"name":"DeepSeek","icon":"DeepSeek.Color"},{"id":9,"name":"Moonshot","icon":"Moonshot"},{"id":88,"name":"Meta","icon":"Meta.Color"},{"id":92,"name":"Nvidia","icon":"Nvidia.Color"},{"id":139,"name":"Flux","icon":"Flux"},{"id":6,"name":"MiniMax","icon":"Minimax.Color"},{"id":8,"name":"Google","icon":"Gemini.Color"},{"id":99,"name":"Others"}]}