AI BENCHY 分类
编程 排名
看看哪些 AI 模型在 编程 上表现最好，哪些更稳定，以及差距主要出现在哪里。 排序方式: 总成本 ↑.
189/189
筛选模型
没有模型匹配当前搜索和筛选条件。
| 排名 | 模型 | 公司 | 编程 得分 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #105 | GPT-5.5 none | OpenAI | 5.5 | 6.3 | $0.231 | 1/3 | 1.35s |
| #55 | GPT-5.6 Luna medium | OpenAI | 5.4 | 7.6 | $0.258 | 1/3 | 10.4s |
| #132 | Claude Sonnet 5 none | Anthropic | 4.6 | 5.7 | $0.287 | 0/3 | 3.67s |
| #60 | Qwen3.6 Flash medium | Qwen | 5.0 | 7.5 | $0.288 | 0/3 | 42.9s |
| #77 | GLM 5.1 medium | Z.ai | 4.6 | 7.1 | $0.288 | 0/3 | 109.6s |
| #37 | Qwen3.6 Plus medium | Qwen | 6.1 | 7.8 | $0.294 | 1/3 | 153.1s |
| #166 | MiniMax M2.5 medium | Minimax | 3.4 | 4.7 | $0.303 | 0/3 | 188.6s |
| #34 | Qwen3.5 Plus 2026-02-15 medium | Qwen | 6.6 | 8.0 | $0.310 | 1/3 | 180.7s |
| #68 | Claude Sonnet 4.6 none | Anthropic | 5.5 | 7.3 | $0.316 | 1/3 | 5.19s |
| #42 | Qwen3.5 Plus 2026-04-20 medium | Qwen | 6.2 | 7.8 | $0.317 | 1/3 | 125.3s |
| #24 | GLM 5 Turbo medium | Z.ai | 8.2 | 8.4 | $0.323 | 2/3 | 45.9s |
| #91 | Mimo V2 PRO medium | Xiaomi | 6.0 | 6.7 | $0.333 | 1/3 | 94.2s |
| #95 | Qwen3.6 27B medium | Qwen | 7.7 | 6.6 | $0.336 | 2/3 | 143.0s |
| #50 | Step 3.7 Flash low | Stepfun | 8.2 | 7.7 | $0.341 | 2/3 | 9.46s |
| #48 | GPT-5.6 Terra low | OpenAI | 6.6 | 7.7 | $0.343 | 1/3 | 9.56s |