Live cost and intelligence for the leading models. Prices are USD per million tokens; blended is a 3:1 input-to-output mix. A 0/0 price means the model is self-hosted (bring your own weights).
Swipe the table sideways for Blended and Score →
| Model | Provider | Input | Output | Blended | Score |
|---|---|---|---|---|---|
| Claude Opus 4.8 | Anthropic | $5 | $25 | $10 | 69.12 |
| Claude Sonnet 5 | Anthropic | $2 | $10 | $4 | 65.92 |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | $2 | 41.47 |
| GPT-5.5 | OpenAI | $5 | $30 | $11.25 | 67.79 |
| GPT-5.4 | OpenAI | $2.50 | $15 | $5.63 | 67.03 |
| GPT-5.1 | OpenAI | $1.25 | $10 | $3.44 | 57.22 |
| Gemini 3.1 Pro | $2 | $12 | $4.50 | 65.17 | |
| Gemini 3 Pro | $2 | $12 | $4.50 | 59.42 | |
| Gemini 3.5 Flash | $1.50 | $9 | $3.38 | 62.52 | |
| Mistral Large | Mistral | $0.50 | $1.50 | $0.75 | — |
| Mistral Medium | Mistral | $1.50 | $7.50 | $3 | — |
Scores are BenchLM’s raw overall score (may be provisional), not the calibrated site ranking. Prices in USD per million tokens. Scores from BenchLM.ai. Rates updated Oct 10, 2026.