{"data":[{"availability":"live","cache_hit_per_million":0.26,"context":128000,"description":"Z.AI flagship from the 5.1 generation. Strong across reasoning, code, and Chinese understanding.","id":"glm-5.1","input_per_million":1.4,"line":"nube","line_code":"N","max_output":32000,"name":"glm-5.1 [N]","object":"model","output_per_million":4.4,"owned_by":"zhipu","tags":[]},{"availability":"live","cache_hit_per_million":0.26,"context":128000,"description":"Z.AI's latest flagship. Open-weights, top-tier reasoning and long-horizon coding.","id":"glm-5.2","input_per_million":1.4,"line":"nube","line_code":"N","max_output":32000,"name":"glm-5.2 [N]","object":"model","output_per_million":4.4,"owned_by":"zhipu","tags":[]},{"availability":"live","cache_hit_per_million":0.1,"context":256000,"description":"Moonshot Kimi K2.5. Long-context general model.","id":"kimi-k2.5","input_per_million":0.6,"line":"nube","line_code":"N","max_output":32000,"name":"kimi-k2.5 [N]","object":"model","output_per_million":3.0,"owned_by":"moonshot","tags":[]},{"availability":"live","cache_hit_per_million":0.16,"context":256000,"description":"Moonshot Kimi K2.6. Long-context general model.","id":"kimi-k2.6","input_per_million":0.95,"line":"nube","line_code":"N","max_output":32000,"name":"kimi-k2.6 [N]","object":"model","output_per_million":4.0,"owned_by":"moonshot","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":1048576,"description":"Moonshot Kimi K3. Flagship open-weight MoE model with native multimodal understanding.","id":"kimi-k3","input_per_million":3.0,"line":"nube","line_code":"N","max_output":32000,"name":"kimi-k3 [N]","object":"model","output_per_million":15.0,"owned_by":"moonshot","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":400000,"description":"OpenAI flagship. Strong general reasoning and tool use. Listed at OpenAI's official price.","id":"gpt-5.5","input_per_million":5.0,"line":"jinwang","line_code":"J","max_output":128000,"name":"gpt-5.5 [J]","object":"model","output_per_million":30.0,"owned_by":"openai","tags":[]},{"availability":"live","cache_hit_per_million":1.5,"context":200000,"description":"Anthropic's top-tier reasoning model. Listed at Anthropic's official price.","id":"claude-opus-4-8","input_per_million":15.0,"line":"jinwang","line_code":"J","max_output":64000,"name":"claude-opus-4-8 [J]","object":"model","output_per_million":75.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":1.5,"context":200000,"description":"Anthropic Opus 4.7. Top-tier reasoning. Listed at Anthropic's official price.","id":"claude-opus-4-7","input_per_million":15.0,"line":"jinwang","line_code":"J","max_output":64000,"name":"claude-opus-4-7 [J]","object":"model","output_per_million":75.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":200000,"description":"Anthropic Sonnet 5. Balanced reasoning at moderate cost. Listed at Anthropic's official price.","id":"claude-sonnet-5","input_per_million":3.0,"line":"jinwang","line_code":"J","max_output":64000,"name":"claude-sonnet-5 [J]","object":"model","output_per_million":15.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":200000,"description":"Anthropic's balanced flagship. Strong reasoning at moderate cost. Listed at Anthropic's official price.","id":"claude-sonnet-4-6","input_per_million":3.0,"line":"jinwang","line_code":"J","max_output":64000,"name":"claude-sonnet-4-6 [J]","object":"model","output_per_million":15.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":1.0,"context":200000,"description":"Anthropic Fable 5. Listed at Anthropic's official price.","id":"claude-fable-5","input_per_million":10.0,"line":"jinwang","line_code":"J","max_output":64000,"name":"claude-fable-5 [J]","object":"model","output_per_million":50.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.0028,"context":128000,"description":"DeepSeek V4 with thinking and non-thinking modes.","id":"deepseek-v4-flash","input_per_million":0.14,"line":"mixai","line_code":"M","max_output":8192,"name":"deepseek-v4-flash-0731 [M]","object":"model","output_per_million":0.28,"owned_by":"deepseek","tags":[]},{"availability":"live","cache_hit_per_million":0.0145,"context":128000,"description":"Advanced DeepSeek V4. Higher-quality reasoning.","id":"deepseek-v4-pro","input_per_million":1.74,"line":"mixai","line_code":"M","max_output":8192,"name":"deepseek-v4-pro-0813 [M]","object":"model","output_per_million":3.48,"owned_by":"deepseek","tags":[]},{"availability":"live","context":400000,"description":"OpenAI flagship. Strong general reasoning and tool use.","id":"gpt-5.5","input_per_million":75.0,"line":"mixai","line_code":"M","max_output":128000,"name":"gpt-5.5 [M]","object":"model","output_per_million":450.0,"owned_by":"openai","tags":[]},{"availability":"live","context":400000,"description":"OpenAI GPT-5.4. High-capability reasoning.","id":"gpt-5.4","input_per_million":75.0,"line":"mixai","line_code":"M","max_output":128000,"name":"gpt-5.4 [M]","object":"model","output_per_million":450.0,"owned_by":"openai","tags":[]},{"availability":"live","cache_hit_per_million":0.075,"context":400000,"description":"Smaller, cheaper GPT-5.4 for high-volume workloads.","id":"gpt-5.4-mini","input_per_million":0.75,"line":"mixai","line_code":"M","max_output":128000,"name":"gpt-5.4-mini [M]","object":"model","output_per_million":4.5,"owned_by":"openai","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":400000,"description":"GPT-5.3 tuned for code generation and agentic coding.","id":"gpt-5.3-codex","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":128000,"name":"gpt-5.3-codex [M]","object":"model","output_per_million":16.0,"owned_by":"openai","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":200000,"description":"Anthropic's top-tier reasoning model.","id":"claude-opus-4-8","input_per_million":5.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-opus-4-8 [M]","object":"model","output_per_million":25.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":200000,"description":"Anthropic Opus 4.7. Top-tier reasoning.","id":"claude-opus-4-7","input_per_million":5.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-opus-4-7 [M]","object":"model","output_per_million":25.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":200000,"description":"Anthropic Opus 4.6. Top-tier reasoning.","id":"claude-opus-4-6","input_per_million":5.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-opus-4-6 [M]","object":"model","output_per_million":25.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":200000,"description":"Anthropic Opus 4.5. Top-tier reasoning.","id":"claude-opus-4-5-20251101","input_per_million":5.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-opus-4-5-20251101 [M]","object":"model","output_per_million":25.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":200000,"description":"Anthropic Sonnet 5. Balanced reasoning at moderate cost.","id":"claude-sonnet-5","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-sonnet-5 [M]","object":"model","output_per_million":10.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":200000,"description":"Anthropic's balanced flagship. Strong reasoning at moderate cost.","id":"claude-sonnet-4-6","input_per_million":3.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-sonnet-4-6 [M]","object":"model","output_per_million":15.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":200000,"description":"Anthropic Sonnet 4.5. Balanced reasoning.","id":"claude-sonnet-4-5-20250929","input_per_million":3.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-sonnet-4-5-20250929 [M]","object":"model","output_per_million":15.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.1,"context":200000,"description":"Fast, low-cost Claude for high-volume tasks.","id":"claude-haiku-4-5-20251001","input_per_million":1.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-haiku-4-5-20251001 [M]","object":"model","output_per_million":5.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":1.0,"context":200000,"description":"Anthropic Fable 5.","id":"claude-fable-5","input_per_million":10.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-fable-5 [M]","object":"model","output_per_million":50.0,"owned_by":"anthropic","tags":[]},{"availability":"live","cache_hit_per_million":0.25,"context":200000,"description":"Anthropic Fable 5.1.","id":"claude-fable-5-1","input_per_million":10.0,"line":"mixai","line_code":"M","max_output":64000,"name":"claude-fable-5-1 [M]","object":"model","output_per_million":50.0,"owned_by":"anthropic","tags":["new"]},{"availability":"live","cache_hit_per_million":0.15,"context":1000000,"description":"Google's fast multimodal model. Low latency.","id":"gemini-3.5-flash","input_per_million":1.5,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3.5-flash [M]","object":"model","output_per_million":9.0,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":1000000,"description":"Google's latest-generation Pro model (preview).","id":"gemini-3.1-pro-preview","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3.1-pro-preview [M]","object":"model","output_per_million":12.0,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":1000000,"description":"Gemini 3.1 Pro with custom tools (preview).","id":"gemini-3.1-pro-preview-customtools","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3.1-pro-preview-customtools [M]","object":"model","output_per_million":12.0,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.025,"context":1000000,"description":"Lightweight, low-cost Gemini 3.1 Flash.","id":"gemini-3.1-flash-lite","input_per_million":0.25,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3.1-flash-lite [M]","object":"model","output_per_million":1.5,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.025,"context":1000000,"description":"Lightweight Gemini 3.1 Flash (preview).","id":"gemini-3.1-flash-lite-preview","input_per_million":0.25,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3.1-flash-lite-preview [M]","object":"model","output_per_million":1.5,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.05,"context":1000000,"description":"Fast Gemini 3 multimodal model (preview).","id":"gemini-3-flash-preview","input_per_million":0.5,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-3-flash-preview [M]","object":"model","output_per_million":3.0,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.31,"context":1000000,"description":"Google's strong reasoning model with 1M context.","id":"gemini-2.5-pro","input_per_million":1.25,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-2.5-pro [M]","object":"model","output_per_million":10.0,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.08,"context":1000000,"description":"Google's fast multimodal model with 1M context.","id":"gemini-2.5-flash","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-2.5-flash [M]","object":"model","output_per_million":2.5,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.025,"context":1000000,"description":"Lightweight, low-cost Gemini 2.5 Flash.","id":"gemini-2.5-flash-lite","input_per_million":0.1,"line":"mixai","line_code":"M","max_output":65536,"name":"gemini-2.5-flash-lite [M]","object":"model","output_per_million":0.4,"owned_by":"google","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"xAI Grok 4.3. Strong reasoning and tool use.","id":"grok-4.3","input_per_million":1.25,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.3 [M]","object":"model","output_per_million":2.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":256000,"description":"xAI Grok 4.2. Strong reasoning and tool use.","id":"grok-4.2","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.2 [M]","object":"model","output_per_million":6.0,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"xAI Grok 4.20 multi-agent variant.","id":"grok-4.20-multi-agent-0309","input_per_million":1.25,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.20-multi-agent-0309 [M]","object":"model","output_per_million":2.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"xAI Grok 4.20, reasoning mode.","id":"grok-4.20-0309-reasoning","input_per_million":1.25,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.20-0309-reasoning [M]","object":"model","output_per_million":2.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"xAI Grok 4.20, non-reasoning mode.","id":"grok-4.20-0309-non-reasoning","input_per_million":1.25,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.20-0309-non-reasoning [M]","object":"model","output_per_million":2.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.5,"context":256000,"description":"xAI Grok 4.1. Strong reasoning and tool use.","id":"grok-4.1","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4.1 [M]","object":"model","output_per_million":6.0,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.05,"context":256000,"description":"xAI Grok 4.1 fast, reasoning mode.","id":"grok-4-1-fast-reasoning","input_per_million":0.2,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4-1-fast-reasoning [M]","object":"model","output_per_million":0.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.05,"context":256000,"description":"xAI Grok 4.1 fast, non-reasoning mode.","id":"grok-4-1-fast-non-reasoning","input_per_million":0.2,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4-1-fast-non-reasoning [M]","object":"model","output_per_million":0.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.05,"context":256000,"description":"xAI Grok 4 fast, reasoning mode.","id":"grok-4-fast-reasoning","input_per_million":0.2,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4-fast-reasoning [M]","object":"model","output_per_million":0.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.05,"context":256000,"description":"xAI Grok 4 fast, non-reasoning mode.","id":"grok-4-fast-non-reasoning","input_per_million":0.2,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4-fast-non-reasoning [M]","object":"model","output_per_million":0.5,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.75,"context":256000,"description":"xAI Grok 4. Flagship reasoning model.","id":"grok-4","input_per_million":3.0,"line":"mixai","line_code":"M","max_output":32000,"name":"grok-4 [M]","object":"model","output_per_million":15.0,"owned_by":"xai","tags":[]},{"availability":"live","cache_hit_per_million":0.3,"context":128000,"description":"Z.AI flagship. Top-tier reasoning and long-horizon coding.","id":"glm-5.2","input_per_million":1.2,"line":"mixai","line_code":"M","max_output":32000,"name":"glm-5.2 [M]","object":"model","output_per_million":4.2,"owned_by":"zhipu","tags":[]},{"availability":"live","cache_hit_per_million":0.18,"context":128000,"description":"Z.AI flagship from the 5.1 generation.","id":"glm-5.1","input_per_million":0.9,"line":"mixai","line_code":"M","max_output":32000,"name":"glm-5.1 [M]","object":"model","output_per_million":4.0,"owned_by":"zhipu","tags":[]},{"availability":"live","cache_hit_per_million":0.03,"context":128000,"description":"DeepSeek V3.2. Efficient general-purpose model.","id":"deepseek-v3.2","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":8192,"name":"deepseek-v3.2 [M]","object":"model","output_per_million":0.45,"owned_by":"deepseek","tags":[]},{"availability":"live","context":128000,"description":"DeepSeek V3 (0324). General-purpose model.","id":"deepseek-v3-0324","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":8192,"name":"deepseek-v3-0324 [M]","object":"model","output_per_million":1.2,"owned_by":"deepseek","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"Moonshot Kimi K2.7, tuned for code.","id":"kimi-k2.7-code","input_per_million":1.0,"line":"mixai","line_code":"M","max_output":32000,"name":"kimi-k2.7-code [M]","object":"model","output_per_million":4.0,"owned_by":"moonshot","tags":[]},{"availability":"live","cache_hit_per_million":0.2,"context":256000,"description":"Moonshot Kimi K2.6. Long-context general model.","id":"kimi-k2.6","input_per_million":1.0,"line":"mixai","line_code":"M","max_output":32000,"name":"kimi-k2.6 [M]","object":"model","output_per_million":4.0,"owned_by":"moonshot","tags":[]},{"availability":"live","cache_hit_per_million":0.12,"context":256000,"description":"Moonshot Kimi K2.5. Long-context general model.","id":"kimi-k2.5","input_per_million":0.6,"line":"mixai","line_code":"M","max_output":32000,"name":"kimi-k2.5 [M]","object":"model","output_per_million":3.0,"owned_by":"moonshot","tags":[]},{"availability":"preview","context":1000000,"description":"Alibaba's newest flagship, a 2.4T-parameter MoE model (95B active). Native text/image/video input. Not yet live on this channel.","id":"qwen3.8-max","input_per_million":2.0,"line":"mixai","line_code":"M","max_output":131072,"name":"qwen3.8-max [M]","object":"model","output_per_million":6.0,"owned_by":"alibaba","tags":["reasoning","tools","day-one"]},{"availability":"preview","context":0,"description":"TypeSafe AI's System One Model. Returns typed, calibrated decisions instead of generated text — built for classification/routing/scoring, not chat. Early access; not yet routable here.","id":"jev","input_per_million":0.0,"line":"mixai","line_code":"M","max_output":0,"name":"Jev [M]","object":"model","output_per_million":0.0,"owned_by":"typesafe","tags":["day-one"]},{"availability":"live","cache_hit_per_million":0.18,"context":256000,"description":"Alibaba Qwen 3.7 Max. Flagship reasoning.","id":"qwen3.7-max","input_per_million":1.8,"line":"mixai","line_code":"M","max_output":32000,"name":"qwen3.7-max [M]","object":"model","output_per_million":5.4,"owned_by":"alibaba","tags":[]},{"availability":"live","cache_hit_per_million":0.06,"context":256000,"description":"Alibaba Qwen 3.7 Plus. Cost-efficient general model.","id":"qwen3.7-plus","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":32000,"name":"qwen3.7-plus [M]","object":"model","output_per_million":1.2,"owned_by":"alibaba","tags":[]},{"availability":"live","cache_hit_per_million":0.14,"context":1000000,"description":"MiniMax M3. General-purpose model.","id":"MiniMax/MiniMax-M3","input_per_million":0.7,"line":"mixai","line_code":"M","max_output":32000,"name":"MiniMax/MiniMax-M3 [M]","object":"model","output_per_million":2.8,"owned_by":"minimax","tags":[]},{"availability":"live","cache_hit_per_million":0.06,"context":1000000,"description":"MiniMax M2.7. General-purpose model.","id":"MiniMax/MiniMax-M2.7","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":32000,"name":"MiniMax/MiniMax-M2.7 [M]","object":"model","output_per_million":12.0,"owned_by":"minimax","tags":[]},{"availability":"live","context":1000000,"description":"MiniMax M2.7. General-purpose model.","id":"MiniMax-M2.7","input_per_million":0.3,"line":"mixai","line_code":"M","max_output":32000,"name":"MiniMax-M2.7 [M]","object":"model","output_per_million":1.2,"owned_by":"minimax","tags":[]},{"availability":"live","context":0,"description":"ByteDance Seedance 2.0 fast, via ArkLin. Text/image-to-video, up to 720p.","id":"doubao-seedance-2-0-fast","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.15,"unit":"video"},"name":"Seedance 2.0 Fast [A]","note":"Via ArkLin (fast). Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"ByteDance Seedance 2.0 standard, via ArkLin. Highest quality, up to 1080p.","id":"doubao-seedance-2.0","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.25,"unit":"video"},"name":"Seedance 2.0 [A]","note":"Via ArkLin. Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Dreamina (即梦) Seedance 2.0 standard, via ArkLin. Up to 1080p.","id":"dreamina-seedance-2-0","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.25,"unit":"video"},"name":"Dreamina Seedance 2.0 [A]","note":"Via ArkLin. Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Dreamina (即梦) Seedance 2.0 fast, via ArkLin. Up to 720p.","id":"dreamina-seedance-2-0-fast","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.15,"unit":"video"},"name":"Dreamina Seedance 2.0 Fast [A]","note":"Via ArkLin (fast). Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Dreamina (即梦) Seedance 2.0 lite, via ArkLin. Lightweight standard tier.","id":"dreamina-seedance-2-0-lite","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.25,"unit":"video"},"name":"Dreamina Seedance 2.0 Lite [A]","note":"Via ArkLin. Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Dreamina (即梦) Seedance 2.0 fast-lite, via ArkLin. Lightweight fast tier.","id":"dreamina-seedance-2-0-fast-lite","input_per_million":0.0,"line":"arklin","line_code":"A","max_output":0,"media":{"per_second_usd":0.15,"unit":"video"},"name":"Dreamina Seedance 2.0 Fast Lite [A]","note":"Via ArkLin (fast). Billed per-token at ArkLin's exact cost (pass-through). $/s is a 720p reserve estimate.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":400000,"description":"OpenAI flagship. Strong general reasoning and tool use.","id":"gpt-5.5","input_per_million":5.0,"line":"arklin","line_code":"A","max_output":128000,"name":"GPT-5.5 [A]","object":"model","output_per_million":30.0,"owned_by":"openai","tags":["reasoning","tools","flagship"]},{"availability":"live","context":400000,"description":"Smaller, cheaper GPT-5 for high-volume workloads.","id":"gpt-5-mini","input_per_million":0.25,"line":"arklin","line_code":"A","max_output":128000,"name":"GPT-5 Mini [A]","object":"model","output_per_million":2.0,"owned_by":"openai","tags":["tools","fast"]},{"availability":"live","context":1000000,"description":"Reliable general-purpose model with large context.","id":"gpt-4.1","input_per_million":2.0,"line":"arklin","line_code":"A","max_output":32000,"name":"GPT-4.1 [A]","object":"model","output_per_million":8.0,"owned_by":"openai","tags":["tools"]},{"availability":"live","context":200000,"description":"Anthropic's top-tier reasoning model.","id":"claude-opus-4-8","input_per_million":5.0,"line":"arklin","line_code":"A","max_output":64000,"name":"Claude Opus 4.8 [A]","object":"model","output_per_million":25.0,"owned_by":"anthropic","tags":["reasoning","premium"]},{"availability":"live","context":200000,"description":"Fast, low-cost Claude for high-volume tasks.","id":"claude-haiku-4-5-20251001","input_per_million":1.0,"line":"arklin","line_code":"A","max_output":64000,"name":"Claude Haiku 4.5 [A]","object":"model","output_per_million":5.0,"owned_by":"anthropic","tags":["fast","tools"]},{"availability":"live","context":1000000,"description":"Google's fast multimodal model. Lowest latency in eval.","id":"gemini-3.5-flash","input_per_million":1.5,"line":"arklin","line_code":"A","max_output":64000,"name":"Gemini 3.5 Flash [A]","object":"model","output_per_million":9.0,"owned_by":"google","tags":["fast","multimodal"]},{"availability":"live","context":1000000,"description":"Google's strong reasoning model with 1M context.","id":"gemini-2.5-pro","input_per_million":1.25,"line":"arklin","line_code":"A","max_output":64000,"name":"Gemini 2.5 Pro [A]","object":"model","output_per_million":10.0,"owned_by":"google","tags":["reasoning","multimodal"]},{"availability":"live","context":1000000,"description":"Google's latest-generation Pro model.","id":"gemini-3.1-pro-preview","input_per_million":2.0,"line":"arklin","line_code":"A","max_output":64000,"name":"Gemini 3.1 Pro (Preview) [A]","object":"model","output_per_million":12.0,"owned_by":"google","tags":["reasoning","multimodal","preview"]},{"availability":"live","context":0,"description":"Whitelisted models. ByteDance text-to-video (mini). Fast, lowest cost, up to 720p.","id":"seedance-2-0-mini","input_per_million":0.0,"line":"ark","line_code":"R","max_output":0,"media":{"per_second_usd":0.081,"unit":"video"},"name":"Seedance 2.0 Mini [R]","note":"Seedance 2.0 Mini. Text/image-to-video: $0.036/s @480p, $0.081/s @720p (caps at 720p). Video-to-video: $0.043/s @480p, $0.096/s @720p. Min duration 5s.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Whitelisted models. ByteDance text-to-video (fast). Balanced speed and quality, up to 720p.","id":"seedance-2-0-fast","input_per_million":0.0,"line":"ark","line_code":"R","max_output":0,"media":{"per_second_usd":0.121,"unit":"video"},"name":"Seedance 2.0 Fast [R]","note":"Seedance 2.0 Fast. Billed per second: $0.121/s (720p). Min duration 5s.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"live","context":0,"description":"Whitelisted models. ByteDance text-to-video (standard). Highest quality, up to 1080p.","id":"seedance-2-0","input_per_million":0.0,"line":"ark","line_code":"R","max_output":0,"media":{"per_second_usd":0.16,"unit":"video"},"name":"Seedance 2.0 [R]","note":"Seedance 2.0. Text/image-to-video: $0.07/s @480p, $0.16/s @720p, $0.40/s @1080p. Video-to-video: from $0.09/s @480p. Min duration 5s.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["video","text-to-video"]},{"availability":"preview","context":0,"description":"ByteDance text-to-image model. High-fidelity generation, synchronous.","id":"doubao-seedream-4-0","input_per_million":0.0,"line":"ark","line_code":"R","max_output":0,"media":{"per_image_usd":0.04,"unit":"image"},"name":"Seedream 4.0 [R]","note":"Text-to-image. Preview: no verified image upstream on the current reseller.","object":"model","output_per_million":0.0,"owned_by":"bytedance","tags":["image","text-to-image"]},{"availability":"live","cache_hit_per_million":0.26,"context":1000000,"description":"Z.AI's newest flagship, focused on complex software engineering and agent tasks. 1M-token context, 128K max output, at the same official rate as GLM-5.2.","id":"glm-5.3","input_per_million":1.4,"line":"glm","line_code":"","max_output":128000,"name":"GLM-5.3","object":"model","output_per_million":4.4,"owned_by":"glm","tags":["reasoning","tools","new"]},{"availability":"live","cache_hit_per_million":0.26,"context":1000000,"description":"Z.AI's latest flagship. 1M context, open-weights, top-tier reasoning and long-horizon coding at a fraction of frontier-model cost.","id":"glm-5.2","input_per_million":1.4,"line":"glm","line_code":"","max_output":16000,"name":"GLM-5.2","object":"model","output_per_million":4.4,"owned_by":"glm","tags":["reasoning","tools","flagship","1m-context"]},{"availability":"live","context":128000,"description":"Z.AI flagship from the 5.1 generation. Strong across reasoning, code, and Chinese understanding.","id":"glm-5.1","input_per_million":1.4,"line":"glm","line_code":"","max_output":16000,"name":"GLM-5.1","object":"model","output_per_million":4.4,"owned_by":"glm","tags":["reasoning","tools"]},{"availability":"live","context":128000,"description":"Z.AI flagship general-purpose model. Strong at reasoning and tool use.","id":"glm-5","input_per_million":1.0,"line":"glm","line_code":"","max_output":16000,"name":"GLM-5","object":"model","output_per_million":3.2,"owned_by":"glm","tags":["reasoning","tools"]},{"availability":"live","context":128000,"description":"Higher-throughput variant of GLM-5 for latency-sensitive applications.","id":"glm-5-turbo","input_per_million":1.2,"line":"glm","line_code":"","max_output":16000,"name":"GLM-5 Turbo","object":"model","output_per_million":4.0,"owned_by":"glm","tags":["reasoning","fast"]},{"availability":"live","context":200000,"description":"Reliable mid-tier model with 200K context. Solid value for production workloads.","id":"glm-4.7","input_per_million":0.6,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.7","object":"model","output_per_million":2.2,"owned_by":"glm","tags":["tools","200k-context"]},{"availability":"live","context":200000,"description":"Free tier of the 4.7 family. Great for prototyping and low-volume use.","id":"glm-4.7-flash","input_per_million":0.0,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.7 Flash","object":"model","output_per_million":0.0,"owned_by":"glm","tags":["free","200k-context"]},{"availability":"live","context":128000,"description":"Previous generation flagship. Same price as 4.7 — prefer 4.7 unless pinning a version.","id":"glm-4.6","input_per_million":0.6,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.6","object":"model","output_per_million":2.2,"owned_by":"glm","tags":["legacy"]},{"availability":"live","context":128000,"description":"4.5-series base model. Stable and widely deployed.","id":"glm-4.5","input_per_million":0.6,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.5","object":"model","output_per_million":2.2,"owned_by":"glm","tags":["stable"]},{"availability":"live","context":128000,"description":"Lightweight 4.5 variant — solid quality at ~3x lower cost than the base model.","id":"glm-4.5-air","input_per_million":0.2,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.5 Air","object":"model","output_per_million":1.1,"owned_by":"glm","tags":["cheap"]},{"availability":"live","context":128000,"description":"Free model from the 4.5 family. Good for testing routes end-to-end.","id":"glm-4.5-flash","input_per_million":0.0,"line":"glm","line_code":"","max_output":16000,"name":"GLM-4.5 Flash","object":"model","output_per_million":0.0,"owned_by":"glm","tags":["free"]},{"availability":"live","cache_hit_per_million":0.003,"context":1000000,"description":"Hybrid reasoning model with both thinking and non-thinking modes. 1M context, 384K max output. Successor id to deepseek-v4-flash (renamed by DeepSeek ~2026-09-10). Off-peak price shown; peak-hour price (Mon-Fri 01:00-04:00 & 06:00-10:00 UTC) is double, see peak field.","id":"deepseek-flash","input_per_million":0.15,"line":"deepseek","line_code":"","max_output":384000,"name":"DeepSeek V4.1 Flash-0910","object":"model","output_per_million":0.6,"owned_by":"deepseek","peak":{"cache_hit_per_million":0.006,"input_per_million":0.3,"output_per_million":1.2},"tags":["reasoning","tools","json","new"]},{"availability":"live","cache_hit_per_million":0.003,"context":1000000,"description":"Hybrid reasoning model with both thinking and non-thinking modes. 1M context, 384K max output. Legacy id, aliased upstream to deepseek-flash (V4.1 Flash). Off-peak price shown; peak-hour price (Mon-Fri 01:00-04:00 & 06:00-10:00 UTC) is double.","id":"deepseek-v4-flash","input_per_million":0.15,"line":"deepseek","line_code":"","max_output":384000,"name":"DeepSeek V4 Flash-0731","object":"model","output_per_million":0.6,"owned_by":"deepseek","peak":{"cache_hit_per_million":0.006,"input_per_million":0.3,"output_per_million":1.2},"tags":["reasoning","tools","json"]},{"availability":"live","cache_hit_per_million":0.003625,"context":1000000,"description":"Advanced V4 variant. Higher quality reasoning. GA release (0813), direct from DeepSeek.","id":"deepseek-v4-pro","input_per_million":0.435,"line":"deepseek","line_code":"","max_output":384000,"name":"DeepSeek V4 Pro-0813","object":"model","output_per_million":0.87,"owned_by":"deepseek","tags":["reasoning","tools","json","new"]},{"availability":"live","cache_hit_per_million":0.0028,"context":1000000,"description":"DeepSeek V4 with thinking and non-thinking modes.","id":"deepseek-v4-flash","input_per_million":0.14,"line":"cm","line_code":"C","max_output":384000,"name":"deepseek-v4-flash-0731 [C]","object":"model","output_per_million":0.28,"owned_by":"deepseek","tags":[]}],"object":"list"}