10,896 prices on 4,759 listings · last collected 17 min ago · 442 flagged changes in 30 days
Every price, as each platform publishes it
A collector reads every public price on eight platforms, daily by default, and keeps the page it read. Nothing is typed in by hand. Search a model to see who charges least, per unit, right now.
Model
Claude 
Gemini 
Qwen 
DeepSeek 
GLM 
Kimi 
Mistral 
MiniMax 
Hunyuan 
Nemotron 
Llama Unit price
All 3,235Per second of output 13Per image 6Per million input tokens 1,107Per million output tokens 1,071Per million tokens 5Per million cached input tokens (read) 594Per million cache-write tokens 224Per thousand characters 2Per second of compute time 12Per run (default parameters) 193Other units (see the platform’s wording) 8
Showing 501–594 of 594
Per million cached input tokens (read)94
| Platform | Model · what it covers | USD | Platform’s own wording | Since |
|---|---|---|---|---|
| OpenRouterOfficial | SpaceXAI: Grok 4.3 (batch)Prompt > 200K | $0.32/M cache read | input_cache_read: 0.00000032 / token (>200000; endpoint xai) | |
| OpenRouterThird-party | DeepSeek: R1 0528via DeepInfra · fp4 | $0.35/M cache read | input_cache_read: 0.00000035 / token (endpoint deepinfra/fp4) | |
| AibermNot disclosed | DeepSeek R1 | $0.359/M cache read | model_ratio 1.56 × $2/M · group default ×0.23 · cache_ratio 0.5 | |
| AibermNot disclosed | Claude Opus 4.5 | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 4.5 Thinking | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 4.6 | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 4.6 Thinking | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 4.7 | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 4.8 | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus 5 | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| AibermNot disclosed | Claude Opus Test | $0.368/M cache read | model_ratio 8 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| OpenRouterOfficial | Google: Gemini 3.1 Pro PreviewPrompt > 200Kvia Google · global | $0.4/M cache read | input_audio_cache: 0.0000004 / token (>200000; endpoint google-vertex/global) | |
| OpenRouterOfficial | Google: Gemini 3.1 Pro PreviewPrompt > 200Kvia Google · global | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint google-vertex/global) | |
| OpenRouterOfficial | Google: Gemini 3.1 Pro Preview Custom ToolsPrompt > 200K | $0.4/M cache read | input_audio_cache: 0.0000004 / token (>200000; endpoint google-ai-studio) | |
| OpenRouterOfficial | Google: Gemini 3.1 Pro Preview Custom ToolsPrompt > 200K | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint google-ai-studio) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 Solvia Amazon Bedrock · global | $0.4/M cache read | input_cache_read: 0.0000004 / token (endpoint amazon-bedrock/global) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 Sol Provia Azure | $0.4/M cache read | input_cache_read: 0.0000004 / token (endpoint azure) | |
| OpenRouterOfficial | OpenAI: GPT-5.6 Sol Pro (batch)Prompt > 272K | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint openai) | |
| OpenRouterOfficial | OpenAI: GPT-5.6 Sol (batch)Prompt > 272K | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint openai) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 TerraPrompt > 272Kvia Azure | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 Terra ProPrompt > 272Kvia Azure | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 SolPrompt > 272Kvia Amazon Bedrock · global | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint amazon-bedrock/global) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 Sol ProPrompt > 272Kvia OpenAI | $0.4/M cache read | input_cache_read: 0.0000004 / token (>272000; endpoint openai) | |
| OpenRouterThird-party | Qwen: Qwen2.5 VL 72B Instruct | $0.4/M cache read | input_cache_read: 0.0000004 / token (endpoint parasail/fp8) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.20Prompt > 200Kvia xAI · zdr | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint xai/zdr) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.20 Multi-AgentPrompt > 200Kvia xAI · zdr | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint xai/zdr) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.3Prompt > 200Kvia xAI · zdr | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint xai/zdr) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.6 (batch) | $0.4/M cache read | input_cache_read: 0.0000004 / token (endpoint xai) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.7 (batch) | $0.4/M cache read | input_cache_read: 0.0000004 / token (endpoint xai) | |
| OpenRouterOfficial | SpaceXAI: Grok Build 0.1Prompt > 200Kvia xAI | $0.4/M cache read | input_cache_read: 0.0000004 / token (>200000; endpoint xai) | |
| RunwareNot disclosed | GPT-5.6 SolPer 1M | $0.4/M cache read | $4.0000000000000003e-7 / cachedInputToken | |
| RunwareNot disclosed | GPT-6 Sol> 272K context | $0.4/M cache read | $4e-7 / cachedInputToken · > 272K context | |
| AibermNot disclosed | Qwen3.8 Max | $0.442/M cache read | model_ratio 4.8 × $2/M · group default ×0.23 · cache_ratio 0.2 | |
| OpenRouterOfficial | Anthropic: Claude Fable 5 (batch) | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint anthropic) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 4.5via Claude Platform on AWS | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 4.6via Claude Platform on AWS | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 4.7via Claude Platform on AWS | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 4.8via Claude Platform on AWS | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 5via Claude Platform on AWS | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | OpenAI: GPT-4.1via Azure | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.4Prompt > 272Kvia Azure | $0.5/M cache read | input_cache_read: 0.0000005 / token (>272000; endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.5via Azure | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint azure) | |
| OpenRouterOfficial | OpenAI: GPT-5.5 (batch)Prompt > 272K | $0.5/M cache read | input_cache_read: 0.0000005 / token (>272000; endpoint openai) | |
| OpenRouterOfficial | OpenAI: GPT-6 Astra Pro (batch) | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint openai) | |
| OpenRouterOfficial | OpenAI: GPT-6 Astra (batch) | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint openai) | |
| OpenRouterOfficial | OpenAI: GPT Chat Latest | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint openai) | |
| OpenRouterOfficial | OpenAI: o3 | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint openai) | |
| OpenRouterOfficial | Qwen: Qwen3.8 Max Prime | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint alibaba) | |
| OpenRouterOfficial | Sakana: Fugu Ultra | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint sakana) | |
| OpenRouterOfficial | Sakana: Fugu Ultra v2 | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint sakana) | |
| OpenRouterOfficial partner | SpaceXAI: Grok 4.6via xAI · zdr | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint xai/zdr) | |
| OpenRouterOfficial partner | SpaceXAI: Grok 4.7via Amazon Bedrock · global | $0.5/M cache read | input_cache_read: 0.0000005 / token (endpoint amazon-bedrock/global) | |
| OpenRouterNot disclosed | xAI: Grok Latest | $0.5/M cache read | input_cache_read: 0.0000005 / token | |
| RunwareNot disclosed | Claude Opus 4.7Per 1M | $0.5/M cache read | $5e-7 / cachedInputToken | |
| RunwareNot disclosed | Claude Opus 4.8Per 1M | $0.5/M cache read | $5e-7 / cachedInputToken | |
| RunwareNot disclosed | Claude Opus 5Per 1M | $0.5/M cache read | $5e-7 / cachedInputToken | |
| RunwareNot disclosed | GPT-5.5Per 1M | $0.5/M cache read | $5e-7 / cachedInputToken | |
| RunwareNot disclosed | Grok 4.7< 200K context | $0.5/M cache read | $5e-7 / cachedInputToken · < 200K context | |
| OpenRouterThird-party | Z.ai: GLM 5.3 Prime | $0.56/M cache read | input_cache_read: 0.00000056 / token (endpoint alibaba) | |
| OpenRouterOfficial partner | Anthropic: Claude Sonnet 4.5Prompt > 200Kvia Claude Platform on AWS | $0.6/M cache read | input_cache_read: 0.0000006 / token (>200000; endpoint claude-on-aws) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.5Prompt > 200Kvia xAI · zdr | $0.6/M cache read | input_cache_read: 0.0000006 / token (>200000; endpoint xai/zdr) | |
| OpenRouterOfficial | Amazon: Nova Premier 1.0 | $0.625/M cache read | input_cache_read: 0.000000625 / token (endpoint amazon-bedrock) | |
| OpenRouterOfficial | OpenAI: GPT-4o (batch) | $0.625/M cache read | input_cache_read: 0.000000625 / token (endpoint openai) | |
| AibermNot disclosed | Claude Fable 5 | $0.736/M cache read | model_ratio 16 × $2/M · group default ×0.23 · cache_ratio 0.1 | |
| OpenRouterOfficial | AionLabs: Aion-3.0 | $0.75/M cache read | input_cache_read: 0.00000075 / token (endpoint aion-labs) | |
| OpenRouterOfficial | AionLabs: Aion 3.5 | $0.75/M cache read | input_cache_read: 0.00000075 / token (endpoint aion-labs) | |
| OpenRouterOfficial | Anthropic: Claude Opus 4.1 (batch) | $0.75/M cache read | input_cache_read: 0.00000075 / token (endpoint anthropic) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 SolPrompt > 272Kvia Amazon Bedrock · global | $0.8/M cache read | input_cache_read: 0.0000008 / token (>272000; endpoint amazon-bedrock/global) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.6 Sol ProPrompt > 272Kvia Azure | $0.8/M cache read | input_cache_read: 0.0000008 / token (>272000; endpoint azure) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.6 (batch)Prompt > 200K | $0.8/M cache read | input_cache_read: 0.0000008 / token (>200000; endpoint xai) | |
| OpenRouterOfficial | SpaceXAI: Grok 4.7 (batch)Prompt > 200K | $0.8/M cache read | input_cache_read: 0.0000008 / token (>200000; endpoint xai) | |
| OpenRouterOfficial partner | Anthropic: Claude Fable 5via Claude Platform on AWS | $1/M cache read | input_cache_read: 0.000001 / token (endpoint claude-on-aws) | |
| OpenRouterOfficial partner | OpenAI: GPT-5.5Prompt > 272Kvia Azure | $1/M cache read | input_cache_read: 0.000001 / token (>272000; endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 Astravia Azure | $1/M cache read | input_cache_read: 0.000001 / token (endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 Astra Provia Azure | $1/M cache read | input_cache_read: 0.000001 / token (endpoint azure) | |
| OpenRouterOfficial | OpenAI: GPT-6 Astra Pro (batch)Prompt > 272K | $1/M cache read | input_cache_read: 0.000001 / token (>272000; endpoint openai) | |
| OpenRouterOfficial | OpenAI: GPT-6 Astra (batch)Prompt > 272K | $1/M cache read | input_cache_read: 0.000001 / token (>272000; endpoint openai) | |
| OpenRouterOfficial | Sakana: Fugu UltraPrompt > 272K | $1/M cache read | input_cache_read: 0.000001 / token (>272000; endpoint sakana) | |
| OpenRouterOfficial | Sakana: Fugu Ultra v2Prompt > 272K | $1/M cache read | input_cache_read: 0.000001 / token (>272000; endpoint sakana) | |
| OpenRouterOfficial partner | SpaceXAI: Grok 4.6Prompt > 200Kvia xAI · zdr | $1/M cache read | input_cache_read: 0.000001 / token (>200000; endpoint xai/zdr) | |
| OpenRouterOfficial partner | SpaceXAI: Grok 4.7Prompt > 200Kvia Amazon Bedrock · global | $1/M cache read | input_cache_read: 0.000001 / token (>200000; endpoint amazon-bedrock/global) | |
| OpenRouterNot disclosed | OpenAI: GPT Astra Latest | $1/M cache read | input_cache_read: 0.000001 / token | |
| OpenRouterNot disclosed | xAI: Grok LatestPrompt > 200K | $1/M cache read | input_cache_read: 0.000001 / token (>200000) | |
| RunwareNot disclosed | Claude Fable 5Per 1M | $1/M cache read | $0.000001 / cachedInputToken | |
| RunwareNot disclosed | GPT-6 Astra≤ 272K context | $1/M cache read | $0.000001 / cachedInputToken · ≤ 272K context | |
| RunwareNot disclosed | Grok 4.7≥ 200K context | $1/M cache read | $0.000001 / cachedInputToken · ≥ 200K context | |
| OpenRouterOfficial partner | OpenAI: GPT-4o (2024-08-06)via Azure | $1.25/M cache read | input_cache_read: 0.00000125 / token (endpoint azure) | |
| OpenRouterOfficial | OpenAI: GPT-4o (2024-11-20) | $1.25/M cache read | input_cache_read: 0.00000125 / token (endpoint openai) | |
| OpenRouterOfficial partner | Anthropic: Claude Opus 4.1 | $1.5/M cache read | input_cache_read: 0.0000015 / token (endpoint amazon-bedrock) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 AstraPrompt > 272Kvia Azure | $2/M cache read | input_cache_read: 0.000002 / token (>272000; endpoint azure) | |
| OpenRouterOfficial partner | OpenAI: GPT-6 Astra ProPrompt > 272Kvia Azure | $2/M cache read | input_cache_read: 0.000002 / token (>272000; endpoint azure) | |
| OpenRouterNot disclosed | OpenAI: GPT Astra LatestPrompt > 272K | $2/M cache read | input_cache_read: 0.000002 / token (>272000) | |
| RunwareNot disclosed | GPT-6 Astra> 272K context | $2/M cache read | $0.000002 / cachedInputToken · > 272K context | |
| OpenRouterThird-party | Magnum v4 72B | $2.5/M cache read | input_cache_read: 0.0000025 / token (endpoint mancer/fp8) |
