未遵循指令 失败
看看哪些 AI 模型最常遇到 未遵循指令，让你在选择前先发现稳定性风险。 排序方式: 失败次数 ↑.
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 | 未遵循指令 次数 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #72 | GPT-5.2 Chat none | OpenAI | 1 | 7.9 | $0.594 | 13/22 | 7.73s |
| #73 | Step 3.7 Flash medium | Stepfun | 1 | 7.9 | $0.495 | 13/22 | 23.5s |
| #74 | Qwen3.8 27B low | Qwen | 1 | 7.9 | ~$0.071 | 15/22 | 39.1s |
| #76 | Qwen3.6 Plus medium | Qwen | 1 | 7.8 | $0.418 | 15/22 | 45.5s |
| #77 | Gemini 3.5 Flash Lite medium | 1 | 7.8 | $0.272 | 14/22 | 5.12s | |
| #84 | Qwen3.7 Flash high | Qwen | 1 | 7.8 | $0.052 | 13/22 | 41.4s |
| #85 | GLM 5 medium | Z.ai | 1 | 7.7 | $0.228 | 15/21 | 33.5s |
| #86 | Seed-2.0-Code low | Bytedance Seed | 1 | 7.7 | $0.767 | 13/22 | 72.2s |
| #87 | Claude Opus 4.6 medium | Anthropic | 1 | 7.7 | $2.961 | 13/22 | 32.2s |
| #90 | GLM 5 Turbo medium | Z.ai | 1 | 7.6 | $0.323 | 14/21 | 23.0s |
| #101 | Mercury 2.5 Preview medium | Inception | 1 | 7.5 | $0.024 | 15/22 | 3.58s |
| #105 | Gemini 3.1 Flash Lite Preview medium | 1 | 7.4 | $0.117 | 14/22 | 4.59s | |
| #107 | Kimi K2.7 Code medium | Moonshot AI | 1 | 7.4 | $0.687 | 11/22 | 80.1s |
| #113 | Claude Sonnet 4.6 none | Anthropic | 1 | 7.3 | $0.661 | 12/22 | 7.67s |
| #114 | GLM 5.3 high | Z.ai | 1 | 7.3 | $0.403 | 13/22 | 27.8s |