AI BENCHY 失败分析
答案错误 失败
看看哪些 AI 模型最常遇到 答案错误，让你在选择前先发现稳定性风险。 排序方式: 失败次数 ↑.
| 排名 | 模型 | 公司 | 答案错误 次数 | 分数 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|
| #117 | Qwen3.5-35B-A3B none | Qwen | 12 | 5.6 | 7/21 | 3.37s |
| #128 | Qwen3.6 Flash none | Qwen | 12 | 5.4 | 7/21 | 1.60s |
| #132 | Mistral Small 4 medium | Mistral | 12 | 5.3 | 5/21 | 9.40s |
| #139 | DeepSeek V4 Flash none | DeepSeek | 12 | 5.0 | 5/21 | 26.8s |
| #151 | Trinity Large Preview none | Arcee AI | 12 | 4.6 | 4/21 | 2.98s |
| #159 | Ling-2.6-1T none | Inclusionai | 12 | 4.3 | 3/21 | 7.72s |
| #108 | Qwen3.5-Flash none | Qwen | 13 | 5.8 | 8/21 | 3.58s |
| #110 | Seed-2.0-Lite none | Bytedance Seed | 13 | 5.8 | 8/21 | 2.49s |
| #112 | GLM 5.1 none | Z.ai | 13 | 5.7 | 7/21 | 4.10s |
| #122 | GLM 4.7 Flash none | Z.ai | 13 | 5.5 | 6/21 | 2.86s |
| #125 | GPT-5.4 none | OpenAI | 13 | 5.5 | 7/21 | 1.42s |
| #131 | Qwen3.5-122B-A10B none | Qwen | 13 | 5.3 | 6/21 | 3.41s |
| #134 | GLM 5 Turbo none | Z.ai | 13 | 5.2 | 6/21 | 2.82s |
| #144 | GPT-5.4 Mini none | OpenAI | 13 | 4.9 | 5/21 | 1.13s |
| #150 | Qwen3 Coder Next medium | Qwen | 13 | 4.6 | 4/21 | 8.58s |