最佳模型
按真实仓库任务、程序合成和命令行智能体基准排名。
| # | 模型 | 均值 | 价格 /1M |
|---|---|---|---|
| 1 | GPT-6 Astra | 89.9 | $30.00 |
| 2 | Gemini 4 Argon | 77.9 | $6.00 |
| 3 | MiniMax M3.1 Flash | 73.7 | $0.375 |
| 4 | Grok 4.7 | 71.0 | $4.00 |
| 5 | Claude Opus 5 | 68.8 | $15.00 |
| 6 | DeepSeek V4-Flash 0731 | 63.8 | — |
| 7 | DeepSeek V4.1-Flash 0910 | 62.6 | $0.750 |
| 8 | GPT 5.6-Sol | 60.4 | $12.00 |
| 9 | DeepSeek V4.1-Flash 1017 | 58.1 | $0.750 |
| 10 | GLM 5.3 | 58.0 | — |
| 11 | Kimi K3 | 57.8 | — |
| 12 | DeepSeek V4-Pro 0813 | 56.9 | $2.64 |
| 13 | MiniMax M3 | 54.0 | $0.750 |
评分基准: DeepSWE v1.1, ProgramBench, NL2Repo-Bench, Terminal-Bench 2.1 · 查看完整榜单