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。
313/313
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 | 答案错误 次数 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #213 | Seed 2.1 Turbo none | Bytedance Seed | 11 | 5.9 | $0.092 | 10/22 | 6.84s |
| #226 | GLM 5V Turbo none | Z.ai | 11 | 5.6 | $0.052 | 8/21 | 2.99s |
| #229 | Mimo V2 PRO none | Xiaomi | 11 | 5.6 | $0.045 | 7/21 | 2.27s |
| #232 | Nemotron 3.5 Lightning high | NVIDIA | 11 | 5.6 | $0.134 | 5/22 | 58.3s |
| #234 | Qwen3.6 27B none | Qwen | 11 | 5.5 | $0.062 | 7/22 | 10.6s |
| #236 | Qwen3.8 27B none | Qwen | 11 | 5.5 | ~$0.006 | 9/22 | 3.12s |
| #245 | Laguna S 2.1 high | Poolside | 11 | 5.4 | $0.114 | 4/22 | 111.6s |
| #248 | Seed-2.0-Code none | Bytedance Seed | 11 | 5.3 | $0.151 | 6/22 | 14.7s |
| #296 | Granite 4.2 8B none | IBM Granite | 11 | 4.3 | $0.026 | 3/22 | 86.3s |
| #115 | KAT-Coder-Pro V2.5 low | Kwaipilot | 10 | 7.3 | $0.400 | 11/22 | 20.2s |
| #122 | KAT-Coder-Pro V2.5 high | Kwaipilot | 10 | 7.2 | $0.506 | 11/22 | 22.1s |
| #135 | GPT-5.5 none | OpenAI | 10 | 7.0 | $0.544 | 12/22 | 2.35s |
| #156 | KAT-Coder-Pro V2.5 none | Kwaipilot | 10 | 6.7 | $0.487 | 11/22 | 26.0s |
| #171 | Laguna XS 2.1 medium | Poolside | 10 | 6.5 | $0.068 | 9/22 | 48.2s |
| #185 | GPT-5.6 Luna low | OpenAI | 10 | 6.2 | $0.051 | 10/22 | 5.20s |