AI BENCHY 分类失败
常识问答：答案错误
常识问答
答案错误
看看哪些 AI 模型在 常识问答 上最容易遇到 答案错误，更快找出薄弱点。 排序方式: 响应时间（平均） ↓.
失败原因
| 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） |
|---|---|---|---|---|---|---|
| #62 | MiMo-V2-Omni medium | Xiaomi | 1 | 3.0 | 0/1 | 234.2s |
| #38 | Gemma 4 26B A4B medium | 1 | 3.0 | 0/1 | 180.9s | |
| #57 | Qwen3.5-35B-A3B medium | Qwen | 1 | 3.0 | 0/1 | 177.4s |
| #140 | Qwen3.5-9B medium | Qwen | 1 | 3.0 | 0/1 | 177.0s |
| #42 | Kimi K2.6 medium | Moonshot AI | 1 | 3.0 | 0/1 | 130.3s |
| #50 | Qwen3.6 Flash medium | Qwen | 1 | 3.0 | 0/1 | 122.9s |
| #36 | Step 3.5 Flash none | Stepfun | 1 | 3.0 | 0/1 | 114.1s |
| #43 | Step 3.5 Flash medium | Stepfun | 1 | 3.0 | 0/1 | 108.4s |
| #12 | Qwen3.5 Plus 2026-02-15 medium | Qwen | 1 | 3.0 | 0/1 | 103.8s |
| #33 | Qwen3.5 Plus 2026-04-20 medium | Qwen | 1 | 3.0 | 0/1 | 92.6s |
| #14 | Gemma 4 31B medium | 1 | 3.0 | 0/1 | 90.1s | |
| #17 | Qwen3.5-27B medium | Qwen | 1 | 3.0 | 0/1 | 85.1s |
| #59 | DeepSeek V3.2 medium | DeepSeek | 1 | 3.0 | 0/1 | 84.0s |
| #69 | Kimi K2.5 medium | Moonshot AI | 1 | 3.0 | 0/1 | 83.9s |
| #37 | MiMo-V2-Pro medium | Xiaomi | 1 | 3.0 | 0/1 | 82.7s |