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: 测试正确 ↑.
443/443
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） |
|---|---|---|---|---|---|---|---|
| #224 | Step 3.7 Flash high | Stepfun | 8 | 0.7 | $1.478 | 11/23 | 72.7s |
| #226 | KAT-Coder-Pro V2.5 low | Kwaipilot | 10 | 0.7 | $0.568 | 11/23 | 19.4s |
| #235 | KAT-Coder-Pro V2.5 high | Kwaipilot | 10 | 0.6 | $0.506 | 11/23 | 21.2s |
| #245 | Beam 1.0 501B A23B low | Reflection | 6 | 0.6 | $0.432 | 11/23 | 4.89s |
| #248 | MiMo-V2.5-Pro medium | Xiaomi | 4 | 0.6 | $0.332 | 11/23 | 56.7s |
| #252 | Qwen3.5-35B-A3B medium | Qwen | 3 | 0.6 | $0.716 | 11/23 | 111.5s |
| #258 | Solar Pro 4 high | Upstage | 6 | 0.6 | $0.077 | 11/23 | 128.9s |
| #259 | Qwen3.5 Plus 2026-02-15 none | Qwen | 11 | 0.6 | $0.176 | 11/23 | 20.1s |
| #263 | DeepSeek V3.2 medium | DeepSeek | 6 | 0.6 | $0.152 | 11/23 | 70.1s |
| #266 | Grok 4.20 medium | X AI | 7 | 0.6 | $1.246 | 11/23 | 36.7s |
| #268 | KAT-Coder-Pro V2.5 medium | Kwaipilot | 9 | 0.6 | $0.669 | 11/23 | 23.9s |
| #272 | Claude Fable 5.1 low | Anthropic | 4 | 0.6 | $3.395 | 11/23 | 11.4s |
| #274 | Beam 1.0 501B A23B high | Reflection | 7 | 0.6 | $0.674 | 11/23 | 9.03s |
| #276 | Apodex 1.1 Mini low | Apodex | 9 | 0.6 | $0.000 | 11/23 | 33.7s |
| #281 | KAT-Coder-Pro V2.5 default | Kwaipilot | 10 | 0.6 | $0.742 | 11/23 | 25.0s |