PowerfulBot
Points per dollar.
One licensed board divided by the blended list price. LMArena and LiveBench stay in separate tables and are never averaged. Blended price is 25% input and 75% output per million tokens. How this is counted.
LMArena rating ÷ blended $ per 1M
LMArena per dollar
99 models with both an LMArena rating and a positive blended list price. CC-BY-4.0 ratings, MIT list prices.
| Model | LMArena | Blended / 1M | Per dollar |
|---|---|---|---|
| Llama 3.1 8B Meta | 1,186.6 | $0.073 | 16,367 |
| Ministral 8B (latest) Mistral | 1,190.9 | $0.15 | 7,940 |
| Mistral Small 3.2 Mistral | 1,338.2 | $0.25 | 5,353 |
| Qwen3.5 Flash Alibaba | 1,396.9 | $0.33 | 4,298 |
| Gemini 2.5 Flash-Lite Google | 1,379.3 | $0.33 | 4,244 |
| GPT-5 Nano OpenAI | 1,319.8 | $0.31 | 4,223 |
| GPT-4.1 nano OpenAI | 1,284.8 | $0.33 | 3,953 |
| GPT-6 Luna OpenAI | 1,389.4 | $0.40 | 3,473 |
| DeepSeek V4 Flash DeepSeek | 1,432 | $0.49 | 2,937 |
| GPT OSS 120B OpenAI | 1,365.6 | $0.49 | 2,801 |
| GPT-4o mini OpenAI | 1,286.4 | $0.49 | 2,639 |
| Llama 3.3 70B Meta | 1,274.3 | $0.74 | 1,722 |
| GPT-5.6 Luna OpenAI | 1,430.5 | $0.95 | 1,506 |
| GPT-5.4 nano OpenAI | 1,372 | $0.99 | 1,389 |
| Qwen Plus Alibaba | 1,326.4 | $1.00 | 1,326 |
| Gemini 3.1 Flash Lite Google | 1,415.8 | $1.19 | 1,192 |
| Mistral Large 3 Mistral | 1,428.5 | $1.25 | 1,143 |
| Qwen3.7 Plus Alibaba | 1,454.5 | $1.30 | 1,119 |
| GPT-4.1 mini OpenAI | 1,340.6 | $1.30 | 1,031 |
| Mistral Large (latest) Mistral | 1,265.3 | $1.25 | 1,012 |
| Aya Expanse 32B Cohere | 1,224 | $1.25 | 979 |
| GPT-3.5-turbo OpenAI | 1,124.9 | $1.25 | 900 |
| Qwen3.5 35B-A3B Alibaba | 1,394.6 | $1.56 | 893 |
| Mistral Medium 3.1 Mistral | 1,424.2 | $1.60 | 890 |
| DeepSeek V4 Pro DeepSeek | 1,451.3 | $1.65 | 880 |
| GPT-5 Mini OpenAI | 1,373.3 | $1.56 | 879 |
| Qwen3-Next 80B-A3B Instruct Alibaba | 1,417 | $1.63 | 872 |
| Qwen3.5 27B Alibaba | 1,408.9 | $1.88 | 751 |
| Gemini 3.5 Flash Lite Google | 1,435 | $1.95 | 736 |
| Gemini 2.5 Flash Google | 1,416.9 | $1.95 | 727 |
| Grok 4.3 xAI | 1,396.8 | $2.19 | 639 |
| Qwen3-VL 235B-A22B Alibaba | 1,420 | $2.28 | 624 |
| Qwen3 235B-A22B Alibaba | 1,419.2 | $2.28 | 624 |
| Gemini 3 Flash Preview Google | 1,465.7 | $2.38 | 617 |
| Qwen3.6 Plus Alibaba | 1,436.3 | $2.38 | 605 |
| Qwen3 32B Alibaba | 1,340.1 | $2.28 | 589 |
| Qwen3.5 122B-A10B Alibaba | 1,417.5 | $2.50 | 567 |
| Qwen3.5 397B-A17B Alibaba | 1,437.5 | $2.85 | 504 |
| Gemini 3.8 Flash Google | 1,498.9 | $3.00 | 500 |
| Gemini 3.7 Flash Google | 1,488.5 | $3.00 | 496 |
| Gemini 3.6 Flash Google | 1,479.5 | $3.00 | 493 |
| Qwen3.8 27B Alibaba | 1,439.7 | $3.20 | 450 |
| Muse Spark 1.3 Meta | 1,490.5 | $3.50 | 426 |
| Muse Spark 1.2 Meta | 1,482.4 | $3.50 | 424 |
| Muse Spark 1.1 Meta | 1,478.3 | $3.50 | 422 |
| Mistral Large 4 Mistral | 1,430.8 | $3.48 | 412 |
| GLM-5.2 Zhipu | 1,470.4 | $3.65 | 403 |
| GPT-5.4 mini OpenAI | 1,411.6 | $3.56 | 396 |
| o4-mini OpenAI | 1,353.4 | $3.58 | 379 |
| o3-mini OpenAI | 1,336.6 | $3.58 | 374 |
| Claude Haiku 4.5 (latest) Anthropic | 1,397 | $4.00 | 349 |
| Qwen3 Max Alibaba | 1,439 | $4.80 | 300 |
| Qwen3.8 Max Alibaba | 1,483.2 | $5.00 | 297 |
| Magistral Medium (latest) Mistral | 1,254.9 | $4.25 | 295 |
| Grok 4.5 xAI | 1,447.7 | $5.00 | 290 |
| Grok 4.6 xAI | 1,428.5 | $5.00 | 286 |
| Grok 4.7 xAI | 1,399 | $5.00 | 280 |
| Qwen Max Alibaba | 1,281.9 | $5.20 | 247 |
| Mistral Medium 3.5 Mistral | 1,420.9 | $6.00 | 237 |
| Qwen3.7 Max Alibaba | 1,474.9 | $6.25 | 236 |
| Qwen3.6 Max Preview Alibaba | 1,446.6 | $6.18 | 234 |
| Qwen3-Coder 480B-A35B Instruct Alibaba | 1,356.5 | $6.00 | 226 |
| o3 OpenAI | 1,410 | $6.50 | 217 |
| GPT-4.1 OpenAI | 1,382.8 | $6.50 | 213 |
| Gemini 3.5 Flash Google | 1,482 | $7.13 | 208 |
| Mistral Medium (latest) Mistral | 1,165.2 | $6.00 | 194 |
| Gemini 4 Argon Google | 1,532.3 | $8.00 | 192 |
| Gemini 2.5 Pro Google | 1,457.7 | $7.81 | 187 |
| GPT-5.1 OpenAI | 1,442.8 | $7.81 | 185 |
| Claude Sonnet 5.5 Anthropic | 1,472.6 | $8.00 | 184 |
| GPT-6.1 Sol OpenAI | 1,443.4 | $8.00 | 180 |
| Claude Sonnet 5 Anthropic | 1,441.9 | $8.00 | 180 |
| GPT-5 OpenAI | 1,406.2 | $7.81 | 180 |
| GPT-6 Sol OpenAI | 1,398.1 | $8.00 | 175 |
| Command A Reasoning Cohere | 1,331 | $8.13 | 164 |
| GPT-4o (2024-11-20) OpenAI | 1,300.5 | $8.13 | 160 |
| Gemini 3.1 Pro Preview Google | 1,480.1 | $9.50 | 156 |
| GPT-5.6 Terra OpenAI | 1,448.1 | $9.50 | 152 |
| Command R+ Cohere | 1,228.9 | $8.13 | 151 |
| GPT-5.2 OpenAI | 1,438.3 | $10.94 | 132 |
| GPT-5.4 OpenAI | 1,469.1 | $11.88 | 124 |
| Claude Sonnet 4.6 Anthropic | 1,457.7 | $12.00 | 121 |
| Gemini Pro Latest Google | 1,148.9 | $9.50 | 121 |
| Claude Sonnet 4.5 (latest) Anthropic | 1,438.5 | $12.00 | 120 |
| Claude Opus 5.5 Anthropic | 1,516.2 | $16.00 | 94.8 |
| Muse Wonder Meta | 1,501.7 | $16.00 | 93.9 |
| GPT-5.6 Sol OpenAI | 1,456.7 | $16.00 | 91.0 |
| Claude Opus 5 Anthropic | 1,508.6 | $20.00 | 75.4 |
| Claude Opus 4.6 Anthropic | 1,501.2 | $20.00 | 75.1 |
| Claude Opus 4.7 Anthropic | 1,489.7 | $20.00 | 74.5 |
| Claude Opus 4.8 Anthropic | 1,460.9 | $20.00 | 73.0 |
| Claude Opus 4.5 (latest) Anthropic | 1,451 | $20.00 | 72.5 |
| GPT-5.5 OpenAI | 1,471.4 | $23.75 | 62.0 |
| GPT-4 Turbo OpenAI | 1,271.7 | $25.00 | 50.9 |
| Claude Fable 5.1 Anthropic | 1,511.9 | $40.00 | 37.8 |
| Claude Fable 5 Anthropic | 1,491.3 | $40.00 | 37.3 |
| GPT-6 Astra OpenAI | 1,441 | $40.00 | 36.0 |
| o1 OpenAI | 1,365.8 | $48.75 | 28.0 |
| GPT-4 OpenAI | 1,206 | $52.50 | 23.0 |
LiveBench per dollar
LiveBench score ÷ blended $ per 1M
2 models where Epoch publishes LiveBench and the list price is positive. This ratio is not added to the LMArena table.
| Model | LiveBench | Blended / 1M | Per dollar |
|---|---|---|---|
| Llama 3.3 70B Meta | 50.2 | $0.74 | 67.8 |
| Sonar Perplexity | 46.9 | $1.00 | 46.9 |