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: 响应时间（平均） ↑.
215/215
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） |
|---|---|---|---|---|---|---|---|
| #127 | gpt-oss-120b medium | OpenAI | 9 | 6.1 | $0.019 | 9/22 | 21.9s |
| #115 | Mimo V2 PRO medium | Xiaomi | 5 | 6.3 | $0.333 | 12/21 | 22.2s |
| #24 | GPT-5.2 medium | OpenAI | 3 | 8.4 | $0.951 | 14/22 | 22.6s |
| #53 | GLM 5 Turbo medium | Z.ai | 4 | 7.6 | $0.323 | 14/21 | 23.0s |
| #98 | GLM 5V Turbo medium | Z.ai | 7 | 6.7 | $0.457 | 11/21 | 23.1s |
| #21 | GPT-5.4 medium | OpenAI | 5 | 8.5 | $1.533 | 15/22 | 23.1s |
| #42 | GLM 5.2 medium | Z.ai | 3 | 7.8 | $0.187 | 15/21 | 23.3s |
| #191 | Grok 4.1 Fast medium | X AI | 4 | 4.7 | $0.069 | 9/19 | 23.8s |
| #85 | KAT-Coder-Pro V2.5 medium | Kwaipilot | 9 | 6.9 | $0.467 | 11/22 | 24.0s |
| #159 | Hy3 preview low | Tencent | 4 | 5.5 | $0.015 | 10/21 | 24.6s |
| #19 | Muse Spark 1.1 medium | Meta | 4 | 8.6 | $1.357 | 15/22 | 25.0s |
| #131 | Qwen3.5-Flash none | Qwen | 13 | 6.1 | $0.073 | 8/22 | 25.3s |
| #97 | KAT-Coder-Pro V2.5 none | Kwaipilot | 10 | 6.7 | $0.476 | 11/22 | 25.6s |
| #44 | Claude Sonnet 4.6 medium | Anthropic | 4 | 7.8 | $2.057 | 14/22 | 25.9s |
| #60 | GPT-5.4 Mini medium | OpenAI | 6 | 7.5 | $0.756 | 12/22 | 25.9s |