Updated 4 November 2026 · Artificial Analysis
AI model API prices compared
What the 48 most capable AI models cost per million tokens, from the top model down. The cheapest capable one is Claude Haiku 5.5 at $0.20.
| # | Model | Intelligence | Price / 1M tokens | Speed | Context |
|---|---|---|---|---|---|
| 1 | Gemini 4 Carbon Google | 60 | $8.00 | 92 tok/s | — |
| 2 | Claude Opus 5.5 Anthropic | 59 | $8.00 | 97 tok/s | 1M |
| 3 | GPT-6 Astra OpenAI | 57 | $20 | 51 tok/s | 1.1M |
| 4 | GPT-6.1 Sol OpenAI | 55 | $4.00 | 62 tok/s | 1.1M |
| 5 | Claude Sonnet 5.5 Anthropic | 55 | $4.00 | 135 tok/s | 1M |
| 6 | GPT-6.1 Terra OpenAI | 51 | $2.00 | 108 tok/s | 1.1M |
| 7 | Gemini 4 Argon Google | 50 | $8.00 | 81 tok/s | 1M |
| 8 | Claude Fable 5.1 Anthropic | 50 | $20 | 52 tok/s | 1M |
| 9 | Kimi K3.1 Pro Kimi · open weights | 47 | $4.00 | 28 tok/s | 1M |
| 10 | GPT-6 Sol OpenAI | 44 | $4.00 | — | 1.1M |
| 11 | Muse Wonder Meta | 44 | $8.00 | 130 tok/s | 1M |
| 12 | Muse Spark 1.3 Meta | 37 | $2.00 | — | 1M |
| 13 | Claude Haiku 5.5 Anthropic | 37 | $0.20 | 236 tok/s | 1M |
| 14 | Grok 4.7 SpaceXAI | 36 | $3.00 | 64 tok/s | 500K |
| 15 | Qwen3.8 Max Alibaba | 34 | $3.00 | 37 tok/s | 1M |
| 16 | DeepSeek V4.1 Flash 1017 DeepSeek · open weights | 33 | $0.53 | 157 tok/s | 1M |
| 17 | GPT-5.6 Terra OpenAI | 32 | $4.50 | — | 1.1M |
| 18 | GLM-5.3 Z AI · open weights | 32 | $2.15 | 63 tok/s | 1M |
| 19 | Grok 4.6 SpaceXAI | 31 | $3.00 | — | 500K |
| 20 | Gemini 3.8 Flash Google | 31 | $1.50 | 116 tok/s | 1M |
| 21 | MiMo-V2.6-Pro Xiaomi · open weights | 30 | $0.54 | 39 tok/s | 1.1M |
| 22 | Ling 3.1 Flash 1019 InclusionAI | 30 | $0.45 | — | — |
| 23 | MiniMax-M3.1-Flash MiniMax · open weights | 29 | $0.26 | 108 tok/s | 1M |
| 24 | Step 5 Preview StepFun · open weights | 29 | $1.43 | 134 tok/s | 1M |
| 25 | GPT-6 Luna OpenAI | 28 | $0.20 | 126 tok/s | 1.1M |
| 26 | GLM-5.3-Flash Z AI · open weights | 28 | $0.24 | 63 tok/s | 1M |
| 27 | MiMo-V2.6-Flash Xiaomi · open weights | 27 | $0.17 | 30 tok/s | 1.1M |
| 28 | Mistral Large 4 Mistral · open weights | 26 | $1.05 | — | 1M |
| 29 | Mistral Large 4 Preview Mistral | 25 | $1.05 | — | — |
| 30 | Muse Spark Open Meta · open weights | 25 | $2.00 | 192 tok/s | 1M |
| 31 | GPT-5.6 Luna OpenAI | 24 | $0.45 | — | 1.1M |
| 32 | DeepSeek V4 Pro 0813 DeepSeek · open weights | 23 | $1.98 | 73 tok/s | 1M |
| 33 | Qwen3.8 27B Alibaba · open weights | 20 | $1.13 | 39 tok/s | 1M |
| 34 | MiniMax-M3 MiniMax · open weights | 19 | $0.53 | 102 tok/s | 1M |
| 35 | Beam 1.0 Reflection | 18 | $0.60 | 321 tok/s | — |
| 36 | Kimi K3 Kimi · open weights | 17 | $6.00 | — | 1M |
| 37 | Nemotron 3.5 Super VL NVIDIA · open weights | 14 | $0.28 | 428 tok/s | 1M |
| 38 | Inkling Small Thinking Machines · open weights | 13 | $0.53 | 201 tok/s | 524K |
| 39 | Inkling Thinking Machines · open weights | 13 | $1.76 | 196 tok/s | 524K |
| 40 | Nemotron 3 Ultra 550B A55B NVIDIA · open weights | 12 | $1.13 | 161 tok/s | 262K |
| 41 | Gemini 3.5 Flash-Lite Google | 12 | $0.85 | 374 tok/s | 1M |
| 42 | Claude 4.5 Haiku Anthropic | 7 | $2.00 | — | — |
| 43 | Muse Glimmer Meta | 6 | $0.53 | 115 tok/s | — |
| 44 | Gemini 3.1 Flash-Lite Google | 6 | $0.56 | — | 1M |
| 45 | Nemotron 3.5 Lightning NVIDIA · open weights | 6 | $0.10 | 271 tok/s | 262K |
| 46 | Granite 4.2 8B IBM · open weights | 5 | $0.11 | 75 tok/s | 131K |
| 47 | Mistral Medium 3.5 Mistral | 5 | $3.00 | 169 tok/s | 262K |
| 48 | Mistral Small 4 Mistral · open weights | 4 | $0.26 | 163 tok/s | 262K |
Blended price is 3 parts input to 1 part output, the cheapest hosted rate tracked by Artificial Analysis. Providers change prices often; check the provider before you build on a number.
Data: Artificial Analysis (intelligence and coding indexes, prices, speed) and OpenRouter (open weights, context windows). Updated 4 November 2026; refreshed every 6 hours.
More rankings
All rankings · Best overall · Best open model · Best value · Cheapest good model · Best for coding · Fastest · Longest context