AI BENCHY 分类失败
编程：答案错误
编程
答案错误
看看哪些 AI 模型在 编程 上最容易遇到 答案错误，更快找出薄弱点。 排序方式: 失败次数 ↑.
| 排名 | 模型 | 公司 | 答案错误 次数 | 分类得分 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|
| #1 | Gemini 3 Flash Preview medium | 1 | 7.9 | 1/2 | 96.0s | |
| #3 | Gemini 3.5 Flash low | 1 | 6.8 | 1/2 | 5.54s | |
| #4 | Gemini 3.1 Pro Preview medium | 1 | 7.0 | 1/2 | 54.3s | |
| #9 | Gemini 3.5 Flash none | 1 | 8.2 | 1/2 | 39.6s | |
| #11 | GPT-5.5 medium | OpenAI | 1 | 8.2 | 1/2 | 69.7s |
| #12 | Gemini 3 Flash Preview low | 1 | 7.3 | 1/2 | 6.66s | |
| #14 | Qwen3.6 Max Preview medium | Qwen | 1 | 8.2 | 1/2 | 178.0s |
| #20 | Qwen3.5 Plus 2026-02-15 medium | Qwen | 1 | 7.6 | 1/2 | 193.8s |
| #21 | Seed-2.0-Lite medium | Bytedance Seed | 1 | 7.0 | 1/2 | 107.7s |
| #25 | Qwen3.5-27B medium | Qwen | 1 | 7.0 | 1/2 | 123.9s |
| #26 | Qwen3.7 Max none | Qwen | 1 | 6.8 | 1/2 | 1.39s |
| #27 | GPT-5.4 medium | OpenAI | 1 | 8.2 | 1/2 | 55.0s |
| #33 | Qwen3.6 Plus medium | Qwen | 1 | 4.1 | 0/2 | 201.7s |
| #34 | Gemini 3.1 Flash Lite Preview medium | 1 | 6.8 | 1/2 | 3.98s | |
| #35 | Gemini 3.1 Flash Lite medium | 1 | 6.8 | 1/2 | 3.59s |