{"data":[{"id":"anthropic/claude-fable-5.1","canonical_slug":"anthropic/claude-fable-5.1","hugging_face_id":null,"name":"anthropic: Claude Fable 5.1","created":1788134400,"description":"Claude Fable 5.1 is Anthropic's model for the most demanding reasoning and long-horizon agentic work, above Opus 5.5 in the Claude lineup. It suits long-running agentic coding, multistep research, and document and spreadsheet work, with a 1M-token context.","context_length":1000000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"claude","instruct_type":null},"pricing":{"completion":"0.00005","image_token":"0.00001","input_cache_read":"0.00000025","input_cache_write":"0.0000125","input_cache_write_1h":"0.00002","internal_reasoning":"0.00005","prompt":"0.00001"},"top_provider":{"context_length":1000000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-06-01","expiration_date":null,"links":{"details":"/v1/models/anthropic/claude-fable-5.1"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"high","supports_max_tokens":false},"pi":{"category":"text","series":"claude-fable","vendor":"anthropic","featured":false,"description_zh":"Claude Fable 5.1 是 Anthropic 面向高难度推理与长程 agent 任务的模型，在 Claude 系列中位于 Opus 5.5 之上。适合长时间运行的 agent 编码、多步研究以及文档与表格工作，上下文 1M token。","providers":["anthropic"],"discount":null}},{"id":"anthropic/claude-haiku-5.5","canonical_slug":"anthropic/claude-haiku-5.5","hugging_face_id":null,"name":"anthropic: Claude Haiku 5.5","created":1791331200,"description":"Claude Haiku 5.5 is Anthropic's fastest current model, built for high-volume, latency-sensitive work such as classification, extraction, routing and subagent tasks. It supports adaptive thinking, tool use, and text, image and PDF input with a 1M-token context and 128K output; prompts over 100,000 tokens are billed at a higher rate.","context_length":1000000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"claude","instruct_type":null},"pricing":{"completion":"0.0000005","image_token":"0.0000001","input_cache_read":"0.00000001","input_cache_write":"0.000000125","input_cache_write_1h":"0.0000002","internal_reasoning":"0.0000005","overrides":[{"completion":"0.0000025","image_token":"0.0000005","input_cache_read":"0.00000005","input_cache_write":"0.000000625","input_cache_write_1h":"0.000001","internal_reasoning":"0.0000025","min_prompt_tokens":100001,"prompt":"0.0000005"}],"prompt":"0.0000001"},"top_provider":{"context_length":1000000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-06-01","expiration_date":null,"links":{"details":"/v1/models/anthropic/claude-haiku-5.5"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"claude-haiku","vendor":"anthropic","featured":false,"description_zh":"Claude Haiku 5.5 是 Anthropic 当前速度最快的模型，面向高并发、对延迟敏感的任务，如分类、抽取、路由和子代理。支持自适应思考、工具调用以及文本、图片和 PDF 输入，上下文 1M token，最多输出 128K；提示超过 10 万 token 时按更高价格计费。","providers":["anthropic","azure"],"discount":null}},{"id":"anthropic/claude-opus-5.5","canonical_slug":"anthropic/claude-opus-5.5","hugging_face_id":null,"name":"anthropic: Claude Opus 5.5","created":1790035200,"description":"Claude Opus 5.5 is built for long-running agentic coding and knowledge work, sitting between Sonnet 5.5 and Fable 5.1 in Anthropic's lineup. Thinking is always on and adaptive, with text, image and PDF input, a 1M-token context and 128K output.","context_length":1000000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"claude","instruct_type":null},"pricing":{"completion":"0.00002","image_token":"0.000004","input_cache_read":"0.0000002","input_cache_write":"0.000005","input_cache_write_1h":"0.000008","internal_reasoning":"0.00002","prompt":"0.000004"},"top_provider":{"context_length":1000000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-06-01","expiration_date":null,"links":{"details":"/v1/models/anthropic/claude-opus-5.5"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"claude-opus","vendor":"anthropic","featured":false,"description_zh":"Claude Opus 5.5 面向长时间运行的 agent 编码与知识工作，在 Anthropic 产品线中介于 Sonnet 5.5 与 Fable 5.1 之间。自适应思考常开，支持文本、图片和 PDF 输入，上下文 1M token，最多输出 128K。","providers":["anthropic","azure"],"discount":null}},{"id":"anthropic/claude-sonnet-5","canonical_slug":"anthropic/claude-sonnet-5","hugging_face_id":null,"name":"anthropic: Claude Sonnet 5","created":1782777600,"description":"Claude Sonnet 5 is the previous Sonnet model, released June 2026 and now marked legacy, with Sonnet 5.5 as its successor. It offers adaptive thinking, text, image and PDF input, a 1M-token context and an earlier January 2026 knowledge cutoff.","context_length":1000000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"claude","instruct_type":null},"pricing":{"completion":"0.00001","image_token":"0.000002","input_cache_read":"0.0000002","input_cache_write":"0.0000025","input_cache_write_1h":"0.000004","internal_reasoning":"0.00001","prompt":"0.000002"},"top_provider":{"context_length":1000000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-01-01","expiration_date":null,"links":{"details":"/v1/models/anthropic/claude-sonnet-5"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"high","supports_max_tokens":false},"pi":{"category":"text","series":"claude-sonnet","vendor":"anthropic","featured":false,"description_zh":"Claude Sonnet 5 是上一代 Sonnet，2026 年 6 月发布，官方已标为旧版，继任者是 Sonnet 5.5。支持自适应思考与文本、图片、PDF 输入，上下文 1M token，知识截止于 2026 年 1 月。","providers":["anthropic","azure"],"discount":null}},{"id":"anthropic/claude-sonnet-5.5","canonical_slug":"anthropic/claude-sonnet-5.5","hugging_face_id":null,"name":"anthropic: Claude Sonnet 5.5","created":1790553600,"description":"Claude Sonnet 5.5 is Anthropic's current Sonnet model, positioned for a balance of speed and intelligence below Opus 5.5. It supports adaptive thinking, tool use, and text, image and PDF input with a 1M-token context and 128K output.","context_length":1000000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"claude","instruct_type":null},"pricing":{"completion":"0.00001","image_token":"0.000002","input_cache_read":"0.0000001","input_cache_write":"0.0000025","input_cache_write_1h":"0.000004","internal_reasoning":"0.00001","prompt":"0.000002"},"top_provider":{"context_length":1000000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-06-01","expiration_date":null,"links":{"details":"/v1/models/anthropic/claude-sonnet-5.5"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"high","supports_max_tokens":false},"pi":{"category":"text","series":"claude-sonnet","vendor":"anthropic","featured":false,"description_zh":"Claude Sonnet 5.5 是 Anthropic 当前的 Sonnet，在速度与智能之间取得平衡，定位低于 Opus 5.5。支持自适应思考、工具调用以及文本、图片和 PDF 输入，上下文 1M token，最多输出 128K。","providers":["anthropic","azure"],"discount":null}},{"id":"bytedance/seed-2.1-lite","canonical_slug":"bytedance/seed-2.1-lite","hugging_face_id":null,"name":"bytedance: Seed 2.1 Lite","created":1789430400,"description":"Seed 2.1 Lite is the low-cost model of ByteDance's Seed 2.1 family, built for high-volume work. It takes text and image input with a 1M-token context, up to 256K output, tool calling, structured output and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"doubao","instruct_type":null},"pricing":{"completion":"0.00000041","image_token":"0.00000012","input_cache_read":"0.000000024","internal_reasoning":"0.00000041","prompt":"0.00000012"},"top_provider":{"context_length":1024000,"max_completion_tokens":256000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/bytedance/seed-2.1-lite"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"seed","vendor":"bytedance","featured":false,"description_zh":"Seed 2.1 Lite 是字节跳动 Seed 2.1 系列的低成本型号，适合高并发的大批量任务。支持文本与图片输入，上下文 1M token，最多输出 256K，支持工具调用、结构化输出与可调的推理强度。","providers":["volcengine"],"discount":null}},{"id":"bytedance/seed-2.1-pro","canonical_slug":"bytedance/seed-2.1-pro","hugging_face_id":null,"name":"bytedance: Seed 2.1 Pro","created":1789430400,"description":"Seed 2.1 Pro is ByteDance's flagship multimodal model for long-running agent work, research and coding. It takes text and image input with a 1M-token context, up to 256K output, tool calling, structured output and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"doubao","instruct_type":null},"pricing":{"completion":"0.0000045","image_token":"0.0000009","input_cache_read":"0.00000018","internal_reasoning":"0.0000045","prompt":"0.0000009"},"top_provider":{"context_length":1024000,"max_completion_tokens":256000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/bytedance/seed-2.1-pro"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"seed","vendor":"bytedance","featured":false,"description_zh":"Seed 2.1 Pro 是字节跳动面向长程 Agent 任务、调研与编码的旗舰多模态模型。支持文本与图片输入，上下文 1M token，最多输出 256K，支持工具调用、结构化输出与可调的推理强度。","providers":["volcengine"],"discount":{"percent_off":20,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.0000036","image_token":"0.00000072","input_cache_read":"0.000000144","internal_reasoning":"0.0000036","prompt":"0.00000072"}}}},{"id":"bytedance/seed-2.1-turbo","canonical_slug":"bytedance/seed-2.1-turbo","hugging_face_id":null,"name":"bytedance: Seed 2.1 Turbo","created":1782604800,"description":"Seed 2.1 Turbo balances quality and cost in ByteDance's Seed 2.1 family, with upgraded coding, agent and multimodal ability. It takes text and image input with a 256K-token context, tool calling, structured output and adjustable reasoning effort.","context_length":256000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"doubao","instruct_type":null},"pricing":{"completion":"0.0000025","image_token":"0.0000005","input_cache_read":"0.0000001","internal_reasoning":"0.0000025","prompt":"0.0000005"},"top_provider":{"context_length":256000,"max_completion_tokens":256000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/bytedance/seed-2.1-turbo"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"seed","vendor":"bytedance","featured":false,"description_zh":"Seed 2.1 Turbo 是字节跳动 Seed 2.1 系列中效果与成本均衡的型号，编码、Agent 与多模态能力全面升级。支持文本与图片输入，上下文 256K token，支持工具调用、结构化输出与可调的推理强度。","providers":["volcengine"],"discount":{"percent_off":20,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000002","image_token":"0.0000004","input_cache_read":"0.00000008","internal_reasoning":"0.000002","prompt":"0.0000004"}}}},{"id":"deepseek/deepseek-v4-flash","canonical_slug":"deepseek/deepseek-v4-flash","hugging_face_id":null,"name":"deepseek: DeepSeek V4 Flash 0731","created":1785456000,"description":"DeepSeek V4 Flash is the fast, low-cost model of the DeepSeek V4 generation, switching between deep thinking and plain chat. Text input with a 1M-token context, up to 384K output, tool calling, structured output and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text-\u003etext","input_modalities":["text"],"output_modalities":["text"],"tokenizer":"deepseek","instruct_type":null},"pricing":{"completion":"0.00000132","input_cache_read":"0.000000014","internal_reasoning":"0.00000132","prompt":"0.00000044"},"top_provider":{"context_length":1024000,"max_completion_tokens":384000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","stop","logprobs","top_logprobs","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/deepseek/deepseek-v4-flash"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"deepseek","vendor":"deepseek","featured":false,"description_zh":"DeepSeek V4 Flash 是 DeepSeek V4 一代中速度快、成本低的型号，可在深度思考与通用对话之间切换。文本输入，上下文 1M token，最多输出 384K，支持工具调用、结构化输出与可调的推理强度。","providers":["volcengine"],"discount":{"percent_off":40,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000000792","input_cache_read":"0.0000000084","internal_reasoning":"0.000000792","prompt":"0.000000264"}}}},{"id":"deepseek/deepseek-v4-pro","canonical_slug":"deepseek/deepseek-v4-pro","hugging_face_id":null,"name":"deepseek: DeepSeek V4 Pro 0813","created":1786579200,"description":"DeepSeek V4 Pro is DeepSeek's flagship hybrid reasoning model, tuned for code generation, mathematical reasoning and agent tool use. Text input with a 1M-token context, up to 384K output, tool calling, structured output and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text-\u003etext","input_modalities":["text"],"output_modalities":["text"],"tokenizer":"deepseek","instruct_type":null},"pricing":{"completion":"0.00000396","input_cache_read":"0.000000044","internal_reasoning":"0.00000396","prompt":"0.00000132"},"top_provider":{"context_length":1024000,"max_completion_tokens":384000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","stop","logprobs","top_logprobs","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/deepseek/deepseek-v4-pro"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"deepseek","vendor":"deepseek","featured":false,"description_zh":"DeepSeek V4 Pro 是深度求索的旗舰混合推理模型，在代码生成、数学推理与 Agent 工具调用上表现突出。文本输入，上下文 1M token，最多输出 384K，支持工具调用、结构化输出与可调的推理强度。","providers":["volcengine"],"discount":{"percent_off":40,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000002376","input_cache_read":"0.0000000264","internal_reasoning":"0.000002376","prompt":"0.000000792"}}}},{"id":"deepseek/deepseek-v4.1-flash","canonical_slug":"deepseek/deepseek-v4.1-flash","hugging_face_id":null,"name":"deepseek: DeepSeek V4.1 Flash","created":1788998400,"description":"DeepSeek V4.1 Flash is the newest DeepSeek hybrid reasoning model, with higher reasoning accuracy and faster responses than V4. It takes text and image input with a 1M-token context, up to 384K output, tool calling, structured output and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"deepseek","instruct_type":null},"pricing":{"completion":"0.0000012","image_token":"0.0000003","input_cache_read":"0.000000006","internal_reasoning":"0.0000012","prompt":"0.0000003"},"top_provider":{"context_length":1024000,"max_completion_tokens":384000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","seed","stop","logprobs","top_logprobs","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/deepseek/deepseek-v4.1-flash"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"deepseek","vendor":"deepseek","featured":false,"description_zh":"DeepSeek V4.1 Flash 是深度求索最新的混合推理模型，推理精度与响应速度较 V4 进一步提升。支持文本与图片输入，上下文 1M token，最多输出 384K，支持工具调用、结构化输出与可调的推理强度。","providers":["azure","volcengine"],"discount":{"percent_off":40,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000072","image_token":"0.00000018","input_cache_read":"0.0000000036","internal_reasoning":"0.00000072","prompt":"0.00000018"}}}},{"id":"google/gemini-3-pro-image","canonical_slug":"google/gemini-3-pro-image","hugging_face_id":null,"name":"google: Nano Banana Pro","created":1779926400,"description":"Nano Banana Pro is Google's reasoning-driven image generation and editing model for professional work, the higher-end counterpart to Nano Banana 2. It is suited to complex graphic design, product mockups and data visualizations that need accurate text rendering.","context_length":65536,"architecture":{"modality":"text+image-\u003eimage","input_modalities":["text","image"],"output_modalities":["image"],"tokenizer":"gemini","instruct_type":null},"pricing":{"completion":"0.000012","image_output":"0.00012","image_token":"0.000002","input_cache_read":"0.0000002","input_cache_write":"0.000000375","internal_reasoning":"0.000012","prompt":"0.000002"},"top_provider":{"context_length":65536,"max_completion_tokens":32768,"is_moderated":false},"per_request_limits":null,"supported_parameters":["max_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3-pro-image"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":[],"supports_max_tokens":false},"pi":{"category":"image","series":"gemini-3","vendor":"google","featured":false,"description_zh":"Nano Banana Pro 是 Google 面向专业场景、具备推理能力的图像生成与编辑模型，是 Nano Banana 2 的高端版本。适合复杂平面设计、产品样机，以及需要准确渲染文字的数据可视化。","providers":["gcp"],"discount":null}},{"id":"google/gemini-3.1-flash-image","canonical_slug":"google/gemini-3.1-flash-image","hugging_face_id":null,"name":"google: Nano Banana 2","created":1779926400,"description":"Nano Banana 2 is Google's high-efficiency image generation and conversational editing model, the faster, lower-cost counterpart to Nano Banana Pro. Built for high-volume use, it outputs images from 0.5K up to 4K and accepts text and image input.","context_length":131072,"architecture":{"modality":"text+image+file+video-\u003eimage","input_modalities":["text","image","file","video"],"output_modalities":["image"],"tokenizer":"gemini","instruct_type":null},"pricing":{"completion":"0.000003","image_output":"0.00006","image_token":"0.0000005","internal_reasoning":"0.000003","prompt":"0.0000005"},"top_provider":{"context_length":131072,"max_completion_tokens":32768,"is_moderated":false},"per_request_limits":null,"supported_parameters":["max_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.1-flash-image"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","minimal"],"default_effort":"minimal","supports_max_tokens":false},"pi":{"category":"image","series":"gemini-3.1","vendor":"google","featured":false,"description_zh":"Nano Banana 2 是 Google 的高效图像生成与对话式编辑模型，比 Nano Banana Pro 更快、成本更低。面向大批量使用，输出分辨率从 0.5K 到 4K，支持文本与图片输入。","providers":["gcp"],"discount":null}},{"id":"google/gemini-3.1-flash-lite-image","canonical_slug":"google/gemini-3.1-flash-lite-image","hugging_face_id":null,"name":"google: Nano Banana 2 Lite","created":1782777600,"description":"Nano Banana 2 Lite is the lowest-latency, lowest-cost model in Google's Nano Banana image family, targeting under 2 seconds end to end. It fits high-volume interactive apps and fast multi-turn edits such as color swaps or background changes, at 1K (1024px) resolution.","context_length":65536,"architecture":{"modality":"text+image+file+video-\u003eimage","input_modalities":["text","image","file","video"],"output_modalities":["image"],"tokenizer":"gemini","instruct_type":null},"pricing":{"completion":"0.0000015","image_output":"0.00003","image_token":"0.00000025","internal_reasoning":"0.0000015","prompt":"0.00000025"},"top_provider":{"context_length":65536,"max_completion_tokens":4096,"is_moderated":false},"per_request_limits":null,"supported_parameters":["max_tokens","temperature","top_p","seed","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.1-flash-lite-image"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","minimal"],"default_effort":"minimal","supports_max_tokens":false},"pi":{"category":"image","series":"gemini-3.1","vendor":"google","featured":false,"description_zh":"Nano Banana 2 Lite 是 Google Nano Banana 图像家族中延迟最低、成本最低的模型，端到端目标 2 秒以内。适合高并发的交互应用与快速多轮编辑，例如换色或换背景，输出 1K（1024px）。","providers":["gcp"],"discount":null}},{"id":"google/gemini-3.5-flash","canonical_slug":"google/gemini-3.5-flash","hugging_face_id":null,"name":"google: Gemini 3.5 Flash","created":1779148800,"description":"Gemini 3.5 Flash is an earlier Flash model in Google's Gemini 3 series, now listed as legacy, aimed at sub-agents, multi-step workflows and fast agentic coding loops. It accepts text, image, audio, video and PDF input with a 1M-token context and up to 65K output tokens.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.0000015","completion":"0.000009","image_token":"0.0000015","input_cache_read":"0.00000015","input_cache_write":"0.0000000833","internal_reasoning":"0.000009","prompt":"0.0000015"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","temperature","top_p","seed","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.5-flash"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low","minimal"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-3.5","vendor":"google","featured":false,"description_zh":"Gemini 3.5 Flash 是 Google Gemini 3 系列中较早的 Flash 模型，官方已列为旧版，面向子 agent、多步工作流与快速的 agent 编码循环。支持文本、图片、音频、视频和 PDF 输入，上下文 1M token，最多输出 65K。","providers":["gcp"],"discount":null}},{"id":"google/gemini-3.5-flash-lite","canonical_slug":"google/gemini-3.5-flash-lite","hugging_face_id":null,"name":"google: Gemini 3.5 Flash Lite","created":1784592000,"description":"Gemini 3.5 Flash Lite is the fastest, most cost-effective model in Google's Gemini 3.5 family. It is designed for high-volume sub-agent tasks, document parsing and simple data extraction, with multimodal input and a 1M-token context.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.0000003","completion":"0.0000025","image_token":"0.0000003","input_cache_read":"0.00000003","input_cache_write":"0.0000000833","internal_reasoning":"0.0000025","prompt":"0.0000003"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","temperature","top_p","seed","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.5-flash-lite"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low","minimal"],"default_effort":"minimal","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-3.5","vendor":"google","featured":false,"description_zh":"Gemini 3.5 Flash Lite 是 Google Gemini 3.5 家族中最快、性价比最高的模型。适合大批量的子 agent 任务、文档解析与简单数据抽取，支持多模态输入，上下文 1M token。","providers":["gcp"],"discount":null}},{"id":"google/gemini-3.6-flash","canonical_slug":"google/gemini-3.6-flash","hugging_face_id":null,"name":"google: Gemini 3.6 Flash","created":1784592000,"description":"Gemini 3.6 Flash is a previous-generation Gemini Flash model that balances speed and multimodal capability for general agentic and everyday tasks. Google highlights code generation, agentic execution and spatial reasoning; it takes text, image, audio, video and PDF input.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.0000015","completion":"0.0000075","image_token":"0.0000015","input_cache_read":"0.00000015","input_cache_write":"0.0000000833","internal_reasoning":"0.0000075","prompt":"0.0000015"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","temperature","top_p","seed","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.6-flash"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low","minimal"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-3.6","vendor":"google","featured":false,"description_zh":"Gemini 3.6 Flash 是上一代 Gemini Flash，兼顾速度与多模态能力，适合通用 agent 与日常任务。官方强调代码生成、agent 执行与空间推理，支持文本、图片、音频、视频和 PDF 输入。","providers":["gcp"],"discount":{"percent_off":50,"ends_at":"2026-12-31T15:59:59Z","uniform":true,"pricing":{"audio":"0.00000075","completion":"0.00000375","image_token":"0.00000075","input_cache_read":"0.000000075","input_cache_write":"0.0000000833","internal_reasoning":"0.00000375","prompt":"0.00000075"}}}},{"id":"google/gemini-3.7-flash","canonical_slug":"google/gemini-3.7-flash","hugging_face_id":null,"name":"google: Gemini 3.7 Flash","created":1786579200,"description":"Gemini 3.7 Flash is the Flash release just before 3.8 in Google's Gemini 3 series, aimed at complex coding, agentic workflows and reliable multi-step execution. It supports low, medium and high thinking levels, multimodal input and a 1M-token context with 65K output.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.0000015","completion":"0.0000075","image_token":"0.0000015","input_cache_read":"0.00000015","input_cache_write":"0.0000000833","internal_reasoning":"0.0000075","prompt":"0.0000015"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","temperature","top_p","seed","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.7-flash"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-3.7","vendor":"google","featured":false,"description_zh":"Gemini 3.7 Flash 是 Gemini 3 系列中 3.8 之前的一版 Flash，面向复杂编码、agent 工作流与可靠的多步执行。支持 low、medium、high 三档思考，多模态输入，上下文 1M token，最多输出 65K。","providers":["gcp"],"discount":{"percent_off":50,"ends_at":"2026-12-31T15:59:59Z","uniform":true,"pricing":{"audio":"0.00000075","completion":"0.00000375","image_token":"0.00000075","input_cache_read":"0.000000075","input_cache_write":"0.0000000833","internal_reasoning":"0.00000375","prompt":"0.00000075"}}}},{"id":"google/gemini-3.8-flash","canonical_slug":"google/gemini-3.8-flash","hugging_face_id":null,"name":"google: Gemini 3.8 Flash","created":1788307200,"description":"Gemini 3.8 Flash is the newest Flash model, which Google calls its most intelligent, built for long-horizon software engineering, autonomous agents and complex enterprise workflows. It supports low, medium and high thinking levels and a 1M-token context with multimodal input.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.0000015","completion":"0.0000075","image_token":"0.0000015","input_cache_read":"0.00000015","input_cache_write":"0.0000000833","internal_reasoning":"0.0000075","prompt":"0.0000015"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","temperature","top_p","seed","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-3.8-flash"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-3.8","vendor":"google","featured":false,"description_zh":"Gemini 3.8 Flash 是最新的 Flash 模型，Google 称其为最智能的 Flash，面向长程软件工程、自主 agent 与复杂的企业工作流。支持 low、medium、high 三档思考，多模态输入，上下文 1M token。","providers":["gcp"],"discount":{"percent_off":50,"ends_at":"2026-12-31T15:59:59Z","uniform":true,"pricing":{"audio":"0.00000075","completion":"0.00000375","image_token":"0.00000075","input_cache_read":"0.000000075","input_cache_write":"0.0000000833","internal_reasoning":"0.00000375","prompt":"0.00000075"}}}},{"id":"google/gemini-4-argon","canonical_slug":"google/gemini-4-argon","hugging_face_id":null,"name":"google: Gemini 4 Argon","created":1791849600,"description":"Gemini 4 Argon is Google's frontier model and the first in the Gemini 4 family, built for deep work on complex real-world tasks in coding, finance, law, cybersecurity and science. It supports low, medium and high thinking levels, multimodal input including PDF, audio and video, and a 1M-token context.","context_length":1048576,"architecture":{"modality":"text+image+file+audio+video-\u003etext","input_modalities":["text","image","file","audio","video"],"output_modalities":["text"],"tokenizer":"gemini","instruct_type":null},"pricing":{"audio":"0.000004","completion":"0.00002","image_token":"0.000004","input_cache_read":"0.0000002","internal_reasoning":"0.00002","prompt":"0.000004"},"top_provider":{"context_length":1048576,"max_completion_tokens":65536,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","stop","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-4-argon"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gemini-4","vendor":"google","featured":false,"description_zh":"Gemini 4 Argon 是 Google 的前沿模型，也是 Gemini 4 系列的首个模型，面向编程、金融、法律、网络安全与科学等领域的复杂实际任务。支持 low、medium、high 三档思考，支持 PDF、音频、视频等多模态输入，上下文 1M token。","providers":["gcp"],"discount":null}},{"id":"google/gemini-nano-banana-2.1","canonical_slug":"google/gemini-nano-banana-2.1","hugging_face_id":null,"name":"google: Nano Banana 2.1","created":1791244800,"description":"Nano Banana 2.1 is Google's upgrade to Nano Banana 2, keeping Flash-level speed and cost while improving visual quality, prompt adherence, multi-turn character consistency and text rendering. It outputs 1K, 2K and 4K images and accepts up to 14 reference images.","context_length":131072,"architecture":{"modality":"text+image+file+video-\u003eimage","input_modalities":["text","image","file","video"],"output_modalities":["image"],"tokenizer":"gemini","instruct_type":null},"pricing":{"completion":"0.0000075","image_output":"0.00003","image_token":"0.0000015","input_cache_read":"0.00000015","internal_reasoning":"0.0000075","prompt":"0.0000015"},"top_provider":{"context_length":131072,"max_completion_tokens":32768,"is_moderated":false},"per_request_limits":null,"supported_parameters":["max_tokens","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/google/gemini-nano-banana-2.1"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["high","medium","minimal"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"image","series":"nano-banana-2","vendor":"google","featured":false,"description_zh":"Nano Banana 2.1 是 Nano Banana 2 的升级版，保持 Flash 级的速度与成本，画质、提示遵循、多轮角色一致性与文字渲染均有提升。输出 1K、2K、4K 图像，最多可参考 14 张图片。","providers":["gcp"],"discount":null}},{"id":"openai/gpt-5.4","canonical_slug":"openai/gpt-5.4","hugging_face_id":null,"name":"openai: GPT-5.4","created":1772668800,"description":"GPT-5.4 is the flagship of OpenAI's GPT-5.4 family, aimed at complex coding and professional work one step below GPT-5.5. It accepts text and image input with a 1.05M-token context, 128K output, and reasoning effort from none to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.000015","image_token":"0.0000025","input_cache_read":"0.00000025","input_cache_write":"0.000003125","internal_reasoning":"0.000015","overrides":[{"completion":"0.0000225","image_token":"0.000005","input_cache_read":"0.0000005","input_cache_write":"0.00000625","internal_reasoning":"0.0000225","min_prompt_tokens":272000,"prompt":"0.000005"}],"prompt":"0.0000025"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2025-08-31","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.4"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","minimal","none"],"default_effort":"none","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.4","vendor":"openai","featured":false,"description_zh":"GPT-5.4 是 OpenAI GPT-5.4 家族的旗舰，面向复杂编码与专业工作，定位比 GPT-5.5 低一档。支持文本与图片输入，上下文 1.05M token，最多输出 128K，推理强度从 none 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.0000105","image_token":"0.00000175","input_cache_read":"0.000000175","input_cache_write":"0.0000021875","internal_reasoning":"0.0000105","overrides":[{"completion":"0.00001575","image_token":"0.0000035","input_cache_read":"0.00000035","input_cache_write":"0.000004375","internal_reasoning":"0.00001575","min_prompt_tokens":272000,"prompt":"0.0000035"}],"prompt":"0.00000175"}}}},{"id":"openai/gpt-5.4-mini","canonical_slug":"openai/gpt-5.4-mini","hugging_face_id":null,"name":"openai: GPT-5.4 Mini","created":1773705600,"description":"GPT-5.4 Mini brings GPT-5.4's strengths to a faster, more efficient model for high-volume coding, computer use and subagent work. It sits between GPT-5.4 and GPT-5.4 Nano, with a 400K-token context, image input and 128K output.","context_length":400000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.0000045","image_token":"0.00000075","input_cache_read":"0.000000075","input_cache_write":"0.0000009375","internal_reasoning":"0.0000045","prompt":"0.00000075"},"top_provider":{"context_length":400000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2025-08-31","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.4-mini"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","minimal","none"],"default_effort":"none","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.4","vendor":"openai","featured":false,"description_zh":"GPT-5.4 Mini 把 GPT-5.4 的能力带到更快、更高效的模型上，适合大批量编码、电脑操作与子 agent 任务。定位介于 GPT-5.4 与 GPT-5.4 Nano 之间，上下文 400K token，支持图片输入，最多输出 128K。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000315","image_token":"0.000000525","input_cache_read":"0.0000000525","input_cache_write":"0.00000065625","internal_reasoning":"0.00000315","prompt":"0.000000525"}}}},{"id":"openai/gpt-5.4-nano","canonical_slug":"openai/gpt-5.4-nano","hugging_face_id":null,"name":"openai: GPT-5.4 Nano","created":1773705600,"description":"GPT-5.4 Nano is the smallest and lowest-priced model in the GPT-5.4 family, built for tasks where speed and cost matter most. Good fits are classification, data extraction, ranking and sub-agents, with a 400K-token context, image input and 128K output.","context_length":400000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00000125","image_token":"0.0000002","input_cache_read":"0.00000002","input_cache_write":"0.00000025","internal_reasoning":"0.00000125","prompt":"0.0000002"},"top_provider":{"context_length":400000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2025-08-31","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.4-nano"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","minimal","none"],"default_effort":"none","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.4","vendor":"openai","featured":false,"description_zh":"GPT-5.4 Nano 是 GPT-5.4 家族中最小、成本最低的模型，面向最看重速度与成本的任务。适合分类、数据抽取、排序与子 agent，上下文 400K token，支持图片输入，最多输出 128K。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000000875","image_token":"0.00000014","input_cache_read":"0.000000014","input_cache_write":"0.000000175","internal_reasoning":"0.000000875","prompt":"0.00000014"}}}},{"id":"openai/gpt-5.5","canonical_slug":"openai/gpt-5.5","hugging_face_id":null,"name":"openai: GPT-5.5","created":1776988800,"description":"GPT-5.5 is OpenAI's flagship for the most complex professional work and coding, positioned above GPT-5.4. It takes text and image input with a 1.05M-token context, 128K output and reasoning effort from none to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00003","image_token":"0.000005","input_cache_read":"0.0000005","input_cache_write":"0.00000625","internal_reasoning":"0.00003","overrides":[{"completion":"0.000045","image_token":"0.00001","input_cache_read":"0.000001","input_cache_write":"0.0000125","internal_reasoning":"0.000045","min_prompt_tokens":272000,"prompt":"0.00001"}],"prompt":"0.000005"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2025-12-01","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.5"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.5","vendor":"openai","featured":false,"description_zh":"GPT-5.5 是 OpenAI 面向最复杂专业工作与编码的旗舰，定位高于 GPT-5.4。支持文本与图片输入，上下文 1.05M token，最多输出 128K，推理强度从 none 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000021","image_token":"0.0000035","input_cache_read":"0.00000035","input_cache_write":"0.000004375","internal_reasoning":"0.000021","overrides":[{"completion":"0.0000315","image_token":"0.000007","input_cache_read":"0.0000007","input_cache_write":"0.00000875","internal_reasoning":"0.0000315","min_prompt_tokens":272000,"prompt":"0.000007"}],"prompt":"0.0000035"}}}},{"id":"openai/gpt-5.6-luna","canonical_slug":"openai/gpt-5.6-luna","hugging_face_id":null,"name":"openai: GPT-5.6 Luna","created":1783555200,"description":"GPT-5.6 Luna is the cost-sensitive model in the GPT-5.6 family, roughly matching the nano models of earlier GPT-5 releases. It suits high-volume workloads and still offers a 1.05M-token context, image input and reasoning effort up to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.0000012","image_token":"0.0000002","input_cache_read":"0.00000002","input_cache_write":"0.00000025","internal_reasoning":"0.0000012","overrides":[{"completion":"0.0000018","image_token":"0.0000004","input_cache_read":"0.00000004","input_cache_write":"0.0000005","internal_reasoning":"0.0000018","min_prompt_tokens":272000,"prompt":"0.0000004"}],"prompt":"0.0000002"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-02-16","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.6-luna"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.6","vendor":"openai","featured":false,"description_zh":"GPT-5.6 Luna 是 GPT-5.6 家族中面向成本敏感场景的模型，大致对应早期 GPT-5 的 nano 档。适合大批量负载，同时保留 1.05M token 上下文、图片输入，推理强度最高到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000084","image_token":"0.00000014","input_cache_read":"0.000000014","input_cache_write":"0.000000175","internal_reasoning":"0.00000084","overrides":[{"completion":"0.00000126","image_token":"0.00000028","input_cache_read":"0.000000028","input_cache_write":"0.00000035","internal_reasoning":"0.00000126","min_prompt_tokens":272000,"prompt":"0.00000028"}],"prompt":"0.00000014"}}}},{"id":"openai/gpt-5.6-sol","canonical_slug":"openai/gpt-5.6-sol","hugging_face_id":null,"name":"openai: GPT-5.6 Sol","created":1783555200,"description":"GPT-5.6 Sol is the flagship of the GPT-5.6 family for complex professional work, comparable to the unsuffixed models of earlier GPT-5 releases. It offers a 1.05M-token context, 128K output, image input and reasoning effort from none to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00002","image_token":"0.000004","input_cache_read":"0.0000004","input_cache_write":"0.000005","internal_reasoning":"0.00002","overrides":[{"completion":"0.00003","image_token":"0.000008","input_cache_read":"0.0000008","input_cache_write":"0.00001","internal_reasoning":"0.00003","min_prompt_tokens":272000,"prompt":"0.000008"}],"prompt":"0.000004"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-02-16","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.6-sol"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.6","vendor":"openai","featured":false,"description_zh":"GPT-5.6 Sol 是 GPT-5.6 家族的旗舰，面向复杂专业工作，大致对应早期 GPT-5 的无后缀主力型号。上下文 1.05M token，最多输出 128K，支持图片输入，推理强度从 none 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000014","image_token":"0.0000028","input_cache_read":"0.00000028","input_cache_write":"0.0000035","internal_reasoning":"0.000014","overrides":[{"completion":"0.000021","image_token":"0.0000056","input_cache_read":"0.00000056","input_cache_write":"0.000007","internal_reasoning":"0.000021","min_prompt_tokens":272000,"prompt":"0.0000056"}],"prompt":"0.0000028"}}}},{"id":"openai/gpt-5.6-terra","canonical_slug":"openai/gpt-5.6-terra","hugging_face_id":null,"name":"openai: GPT-5.6 Terra","created":1783555200,"description":"GPT-5.6 Terra is the mid-range GPT-5.6 model, balancing intelligence and cost and roughly matching the mini models of earlier GPT-5 releases. It sits between Sol and Luna, with a 1.05M-token context, image input and reasoning effort from low to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.000012","image_token":"0.000002","input_cache_read":"0.0000002","input_cache_write":"0.0000025","internal_reasoning":"0.000012","overrides":[{"completion":"0.000018","image_token":"0.000004","input_cache_read":"0.0000004","input_cache_write":"0.000005","internal_reasoning":"0.000018","min_prompt_tokens":272000,"prompt":"0.000004"}],"prompt":"0.000002"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-02-16","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-5.6-terra"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-5.6","vendor":"openai","featured":false,"description_zh":"GPT-5.6 Terra 是 GPT-5.6 的中档模型，在智能与成本之间取得平衡，大致对应早期 GPT-5 的 mini 档。定位介于 Sol 与 Luna 之间，上下文 1.05M token，支持图片输入，推理强度从 low 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.0000084","image_token":"0.0000014","input_cache_read":"0.00000014","input_cache_write":"0.00000175","internal_reasoning":"0.0000084","overrides":[{"completion":"0.0000126","image_token":"0.0000028","input_cache_read":"0.00000028","input_cache_write":"0.0000035","internal_reasoning":"0.0000126","min_prompt_tokens":272000,"prompt":"0.0000028"}],"prompt":"0.0000014"}}}},{"id":"openai/gpt-6-astra","canonical_slug":"openai/gpt-6-astra","hugging_face_id":null,"name":"openai: GPT-6 Astra","created":1788393600,"description":"GPT-6 Astra is the most capable GPT-6 model, made for the most demanding work. Use it for complex reasoning, coding, computer use, research and document creation, with a 1.05M-token context and reasoning effort from low to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00005","image_token":"0.00001","input_cache_read":"0.000001","input_cache_write":"0.0000125","internal_reasoning":"0.00005","overrides":[{"completion":"0.000075","image_token":"0.00002","input_cache_read":"0.000002","input_cache_write":"0.000025","internal_reasoning":"0.000075","min_prompt_tokens":272000,"prompt":"0.00002"}],"prompt":"0.00001"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-04-30","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-6-astra"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-6","vendor":"openai","featured":false,"description_zh":"GPT-6 Astra 是 GPT-6 中能力最强的模型，面向要求最高的工作。适合复杂推理、编码、电脑操作、研究与文档创作，上下文 1.05M token，推理强度从 low 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000035","image_token":"0.000007","input_cache_read":"0.0000007","input_cache_write":"0.00000875","internal_reasoning":"0.000035","overrides":[{"completion":"0.0000525","image_token":"0.000014","input_cache_read":"0.0000014","input_cache_write":"0.0000175","internal_reasoning":"0.0000525","min_prompt_tokens":272000,"prompt":"0.000014"}],"prompt":"0.000007"}}}},{"id":"openai/gpt-6-luna","canonical_slug":"openai/gpt-6-luna","hugging_face_id":null,"name":"openai: GPT-6 Luna","created":1790035200,"description":"GPT-6 Luna is the most efficient GPT-6 model, made for focused, high-volume tasks. It keeps a 1.05M-token context, image input, 128K output and reasoning effort from none to xhigh, so it handles long inputs at low cost.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.0000005","image_token":"0.0000001","input_cache_read":"0.00000001","input_cache_write":"0.000000125","internal_reasoning":"0.0000005","overrides":[{"completion":"0.00000075","image_token":"0.0000002","input_cache_read":"0.00000002","input_cache_write":"0.00000025","internal_reasoning":"0.00000075","min_prompt_tokens":272000,"prompt":"0.0000002"}],"prompt":"0.0000001"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-05-18","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-6-luna"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-6","vendor":"openai","featured":false,"description_zh":"GPT-6 Luna 是 GPT-6 中效率最高的模型，面向目标明确的大批量任务。保留 1.05M token 上下文、图片输入、最多 128K 输出，推理强度从 none 到 xhigh，能以低成本处理长输入。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000035","image_token":"0.00000007","input_cache_read":"0.000000007","input_cache_write":"0.0000000875","internal_reasoning":"0.00000035","overrides":[{"completion":"0.000000525","image_token":"0.00000014","input_cache_read":"0.000000014","input_cache_write":"0.000000175","internal_reasoning":"0.000000525","min_prompt_tokens":272000,"prompt":"0.00000014"}],"prompt":"0.00000007"}}}},{"id":"openai/gpt-6-sol","canonical_slug":"openai/gpt-6-sol","hugging_face_id":null,"name":"openai: GPT-6 Sol","created":1790035200,"description":"GPT-6 Sol is the GPT-6 model built for complex coding and agentic workflows, between Luna and Astra in the family. It has a 1.05M-token context, image input, 128K output and reasoning effort from none to xhigh; GPT-6.1 Sol is its newer version.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00001","image_token":"0.000002","input_cache_read":"0.0000002","input_cache_write":"0.0000025","internal_reasoning":"0.00001","overrides":[{"completion":"0.000015","image_token":"0.000004","input_cache_read":"0.0000004","input_cache_write":"0.000005","internal_reasoning":"0.000015","min_prompt_tokens":272000,"prompt":"0.000004"}],"prompt":"0.000002"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-04-20","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-6-sol"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-6","vendor":"openai","featured":false,"description_zh":"GPT-6 Sol 是 GPT-6 中面向复杂编码与 agent 工作流的模型，在家族中介于 Luna 与 Astra 之间。上下文 1.05M token，支持图片输入，最多输出 128K，推理强度从 none 到 xhigh；GPT-6.1 Sol 是它的新版本。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000007","image_token":"0.0000014","input_cache_read":"0.00000014","input_cache_write":"0.00000175","internal_reasoning":"0.000007","overrides":[{"completion":"0.0000105","image_token":"0.0000028","input_cache_read":"0.00000028","input_cache_write":"0.0000035","internal_reasoning":"0.0000105","min_prompt_tokens":272000,"prompt":"0.0000028"}],"prompt":"0.0000014"}}}},{"id":"openai/gpt-6.1-sol","canonical_slug":"openai/gpt-6.1-sol","hugging_face_id":null,"name":"openai: GPT-6.1 Sol","created":1790640000,"description":"GPT-6.1 Sol is the newer Sol model, which OpenAI says delivers near-Astra performance at lower cost for complex coding, computer use and professional work. It has a 1.05M-token context, image input and reasoning effort from low to xhigh.","context_length":1050000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"gpt","instruct_type":null},"pricing":{"completion":"0.00001","image_token":"0.000002","input_cache_read":"0.0000001","input_cache_write":"0.0000025","internal_reasoning":"0.00001","overrides":[{"completion":"0.000015","image_token":"0.000004","input_cache_read":"0.0000002","input_cache_write":"0.000005","internal_reasoning":"0.000015","min_prompt_tokens":272000,"prompt":"0.000004"}],"prompt":"0.000002"},"top_provider":{"context_length":1050000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","response_format","structured_outputs","max_tokens","max_completion_tokens","seed","reasoning","include_reasoning","stream","n","verbosity"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":"2026-04-30","expiration_date":null,"links":{"details":"/v1/models/openai/gpt-6.1-sol"},"reasoning":{"mandatory":false,"default_enabled":false,"supported_efforts":["xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"gpt-6.1","vendor":"openai","featured":false,"description_zh":"GPT-6.1 Sol 是较新的 Sol 模型，OpenAI 称它以更低成本提供接近 Astra 的表现，适合复杂编码、电脑操作与专业工作。上下文 1.05M token，支持图片输入，推理强度从 low 到 xhigh。","providers":["azure"],"discount":{"percent_off":30,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.000007","image_token":"0.0000014","input_cache_read":"0.00000007","input_cache_write":"0.00000175","internal_reasoning":"0.000007","overrides":[{"completion":"0.0000105","image_token":"0.0000028","input_cache_read":"0.00000014","input_cache_write":"0.0000035","internal_reasoning":"0.0000105","min_prompt_tokens":272000,"prompt":"0.0000028"}],"prompt":"0.0000014"}}}},{"id":"z-ai/glm-5.2","canonical_slug":"z-ai/glm-5.2","hugging_face_id":null,"name":"zai: GLM 5.2","created":1781654400,"description":"GLM 5.2 is Z.ai's model for complex development work, with clear gains in coding, long-context understanding and long-running tasks. Text input with a 1M-token context, up to 128K output, tool calling and adjustable reasoning effort.","context_length":1024000,"architecture":{"modality":"text-\u003etext","input_modalities":["text"],"output_modalities":["text"],"tokenizer":"glm","instruct_type":null},"pricing":{"completion":"0.0000044","input_cache_read":"0.00000026","internal_reasoning":"0.0000044","prompt":"0.0000014"},"top_provider":{"context_length":1024000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","max_completion_tokens","temperature","top_p","seed","stop","logprobs","top_logprobs","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/z-ai/glm-5.2"},"reasoning":{"mandatory":false,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low","minimal","none"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"glm","vendor":"z-ai","featured":false,"description_zh":"GLM 5.2 是智谱面向复杂开发场景的模型，在编码、长上下文理解与长程任务上有明显提升。文本输入，上下文 1M token，最多输出 128K，支持工具调用与可调的推理强度。","providers":["volcengine"],"discount":{"percent_off":40,"ends_at":"2026-10-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000264","input_cache_read":"0.000000156","internal_reasoning":"0.00000264","prompt":"0.00000084"}}}},{"id":"z-ai/glm-5.3","canonical_slug":"z-ai/glm-5.3","hugging_face_id":null,"name":"zai: GLM 5.3","created":1787011200,"description":"GLM 5.3 is a text model from Z.ai in the GLM 5 family. Text input with a 1M-token context, up to 128K output, tool calling and structured output, and it always reasons before it answers.","context_length":1048576,"architecture":{"modality":"text-\u003etext","input_modalities":["text"],"output_modalities":["text"],"tokenizer":"glm","instruct_type":null},"pricing":{"completion":"0.0000044","input_cache_read":"0.00000026","internal_reasoning":"0.0000044","prompt":"0.0000014"},"top_provider":{"context_length":1048576,"max_completion_tokens":131072,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","response_format","structured_outputs","max_tokens","max_completion_tokens","temperature","top_p","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/z-ai/glm-5.3"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"glm","vendor":"z-ai","featured":false,"description_zh":"GLM 5.3 是智谱 GLM 5 系列的文本模型。文本输入，上下文 1M token，最多输出 128K，支持工具调用与结构化输出，回答前始终先思考。","providers":["siliconflow"],"discount":{"percent_off":30,"ends_at":"2026-12-31T15:59:59Z","uniform":true,"pricing":{"completion":"0.00000308","input_cache_read":"0.000000182","internal_reasoning":"0.00000308","prompt":"0.00000098"}}}},{"id":"z-ai/glm-5.3-flash","canonical_slug":"z-ai/glm-5.3-flash","hugging_face_id":null,"name":"zai: GLM 5.3 Flash","created":1787875200,"description":"GLM 5.3 Flash is the first natively multimodal model of the GLM 5 family, built on a low-cost architecture. It takes text and image input with a 1M-token context, up to 128K output and tool calling, and always reasons before it answers.","context_length":1024000,"architecture":{"modality":"text+image+file-\u003etext","input_modalities":["text","image","file"],"output_modalities":["text"],"tokenizer":"glm","instruct_type":null},"pricing":{"completion":"0.0000005","image_token":"0.00000015","input_cache_read":"0.00000003","internal_reasoning":"0.0000005","prompt":"0.00000015"},"top_provider":{"context_length":1024000,"max_completion_tokens":128000,"is_moderated":false},"per_request_limits":null,"supported_parameters":["tools","tool_choice","parallel_tool_calls","max_tokens","max_completion_tokens","temperature","top_p","seed","stop","logprobs","top_logprobs","reasoning","include_reasoning","stream"],"default_parameters":{},"supported_voices":null,"knowledge_cutoff":null,"expiration_date":null,"links":{"details":"/v1/models/z-ai/glm-5.3-flash"},"reasoning":{"mandatory":true,"default_enabled":true,"supported_efforts":["max","xhigh","high","medium","low"],"default_effort":"medium","supports_max_tokens":false},"pi":{"category":"text","series":"glm","vendor":"z-ai","featured":false,"description_zh":"GLM 5.3 Flash 是 GLM 5 系列首个原生多模态模型，采用低成本架构。支持文本与图片输入，上下文 1M token，最多输出 128K，支持工具调用，回答前始终先思考。","providers":["siliconflow","volcengine"],"discount":{"percent_off":30,"ends_at":"2026-12-31T15:59:59Z","uniform":false,"pricing":{"completion":"0.00000035","image_token":"0.000000105","input_cache_read":"0.000000021","internal_reasoning":"0.00000035","prompt":"0.000000105"}}}}]}