⌨️ Coding model board
SWE-bench Verified (higher is better) next to blended API price (3× input + 1× output) / 1M tokens. Official prices preferred when several sources exist. AA index and output speed are from Artificial Analysis when available. This is not OpenRouter usage share.
| # | Model | SWE | AA | Blended $/1M | SWE / $ | tok/s | Source |
|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 4.7Anthropic | 83.47 | 40.7 | $10 | 8.35 | 53 | Official |
| 2 | GPT-5.5OpenAI | 80.58 | 38.4 | $11.25 | 7.16 | 99 | Official |
| 3 | Claude Opus 4.5Anthropic | 79.2 | 23.7 | $10 | 7.92 | 48 | Official |
| 4 | Claude Opus 4.6Anthropic | 78.72 | 26.4 | $10 | 7.87 | 41 | Official |
| 5 | GLM-5.2Zhipu (Z.ai) | 78.7 | 33.7 | $1.94 | 40.56 | 88 | Official |
| 6 | DeepSeek V4 Pro 0813 (deepseek-v4-pro)DeepSeek | 77.64 | 36 | $2.01 | 38.53 | 83 | Official |
| 7 | Qwen3.7-MaxAlibaba (Qwen) | 77.27 | 29.5 | $2.69 | 28.76 | 197 | Official |
| 8 | GPT-5.4OpenAI | 76.86 | 39 | $5.63 | 13.66 | 99 | Official |
| 9 | Claude Sonnet 4Anthropic | 76.8 | — | $6 | 12.8 | — | Official |
| 10 | Kimi K2.6Moonshot (Kimi) | 76.65 | 27 | $1.74 | 44.18 | 45 | Official |
| 11 | Qwen3.6-Max-PreviewAlibaba (Qwen) | 76.65 | 28.4 | $3.02 | 25.36 | 64 | Official |
| 12 | MiniMax-M2.5MiniMax | 75.8 | 22.8 | $0.549 | 138 | 70 | Official |
| 13 | Gemini 3 Flash PreviewGoogle | 75.8 | 17.9 | $1.13 | 67.38 | 209 | Official |
| 14 | Gemini 3.1 Pro PreviewGoogle | 75.62 | 29.7 | $4.5 | 16.8 | 125 | Official |
| 15 | Claude Sonnet 4.6Anthropic | 75.21 | 24.7 | $6 | 12.53 | 45 | Official |
| 16 | Claude Sonnet 4.5Anthropic | 74.8 | — | $6 | 12.47 | — | Official |
| 17 | GPT-5.3 CodexOpenAI | 74.79 | 32.5 | $4.81 | 15.54 | 90 | Official |
| 18 | GPT-5OpenAI | 74.4 | 23 | $3.44 | 21.64 | 109 | Official |
| 19 | GLM-5.1Zhipu (Z.ai) | 74.17 | 26.1 | $1.57 | 47.33 | 55 | Official |
| 20 | Kimi K2.5Moonshot (Kimi) | 73.76 | 23.5 | $0.993 | 74.32 | 57 | OpenRouter |
| 21 | GPT-5.2OpenAI | 73.76 | 30.4 | $4.81 | 15.33 | 83 | Official |
| 22 | Claude Opus 4.1Anthropic | 73.35 | — | $30 | 2.44 | — | Official |
| 23 | GLM 5Zhipu (Z.ai) | 72.8 | 27.9 | $0.93 | 78.28 | 75 | OpenRouter |
| 24 | Kimi K2 0711Moonshot (Kimi) | 71.2 | 12.7 | $1 | 71.02 | 57 | OpenRouter |
| 25 | Claude Opus 4Anthropic | 70.66 | — | $30 | 2.36 | — | Official |
| 26 | DeepSeek-V3.2DeepSeek | 70 | 8.5 | $0.29 | 241 | 41 | Platform |
| 27 | Qwen3-Coder-480B-A35BAlibaba (Qwen) | 69.6 | 11.9 | $1.57 | 44.41 | 60 | Official |
| 28 | GLM 4.6Zhipu (Z.ai) | 68.2 | 14.9 | $0.875 | 77.94 | 36 | OpenRouter |
| 29 | GPT-5.1OpenAI | 67.98 | 24.7 | $3.44 | 19.78 | 119 | Official |
| 30 | Claude Haiku 4.5Anthropic | 66.6 | — | $2 | 33.3 | — | Official |
| 31 | GPT-5 miniOpenAI | 64.67 | 16.8 | $0.688 | 94.07 | 163 | Official |
| 32 | o4-miniOpenAI | 64.6 | 16.7 | $1.93 | 33.56 | 161 | Official |
| 33 | o3OpenAI | 62.32 | 20.2 | $3.5 | 17.81 | 129 | Official |
| 34 | Qwen3.6-PlusAlibaba (Qwen) | 57.85 | 27 | $0.672 | 86.13 | 55 | Official |
| 35 | Gemini 2.5 ProGoogle | 57.56 | 16.1 | $3.44 | 16.74 | 121 | Official |
| 36 | Qwen3 Coder 30B A3B InstructAlibaba (Qwen) | 51.6 | 9.6 | $0.122 | 421 | 76 | OpenRouter |
| 37 | GPT-4.1OpenAI | 48.54 | 12.7 | $3.5 | 13.87 | 150 | Official |
| 38 | o3-miniOpenAI | 42.4 | 12.5 | $1.93 | 22.03 | 210 | Official |
| 39 | DeepSeek V3 0324DeepSeek | 42 | 9.7 | $0.438 | 96 | 76 | OpenRouter |
| 40 | GPT-5 nanoOpenAI | 34.8 | 13 | $0.138 | 253 | 177 | Official |
| 41 | GPT-4oOpenAI | 30.99 | 8.4 | $4.38 | 7.08 | 168 | Official |
| 42 | Gemini 2.5 FlashGoogle | 28.73 | 9.9 | $0.85 | 33.8 | 194 | Official |
| 43 | gpt-oss-120BOpenAI | 26 | 11.6 | $0.263 | 99.05 | 233 | Platform |
| 44 | GPT-4.1 miniOpenAI | 23.94 | 10.2 | $0.7 | 34.2 | 161 | Official |
Prices checked 2026-10-22. SWE scores joined from site benchmarks (max per model). OpenRouter usage rankings are still a separate gap (MOD-12).