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: 响应时间（平均） ↓.
293/293
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） |
|---|---|---|---|---|---|---|---|
| #311 | Step 5 Preview low | Stepfun | 1 | 0.6 | $4.110 | 10/23 | 138.3s |
| #279 | Step 5 Preview medium | Stepfun | 2 | 0.6 | $4.136 | 10/23 | 136.1s |
| #241 | Solar Pro 4 medium | Upstage | 2 | 0.6 | $0.098 | 12/23 | 133.9s |
| #258 | Solar Pro 4 high | Upstage | 3 | 0.6 | $0.077 | 11/23 | 128.9s |
| #244 | Kimi K2.5 medium | Moonshot AI | 2 | 0.6 | $0.841 | 10/23 | 125.8s |
| #150 | Seed-2.0-Code high | Bytedance Seed | 1 | 0.7 | $1.448 | 12/23 | 124.9s |
| #212 | Grok 4.7 high | X AI | 2 | 0.7 | $3.037 | 13/23 | 122.2s |
| #201 | Kimi K2.6 medium | Moonshot AI | 2 | 0.7 | $1.594 | 12/23 | 121.3s |
| #122 | Qwen3.8 2.4T A95B high | Qwen | 2 | 0.8 | $2.949 | 15/23 | 120.9s |
| #89 | DeepSeek V4 Flash 0731 high | DeepSeek | 2 | 0.8 | $0.576 | 15/23 | 118.8s |
| #163 | Grok 4.7 medium | X AI | 1 | 0.7 | $2.938 | 15/23 | 117.7s |
| #222 | LongCat 2.0 low | Meituan | 1 | 0.7 | $0.630 | 10/23 | 115.0s |
| #116 | Qwen3.5-27B medium | Qwen | 2 | 0.8 | $1.135 | 14/23 | 114.7s |
| #358 | Laguna S 2.1 high | Poolside | 1 | 0.5 | $0.145 | 4/23 | 113.6s |
| #252 | Qwen3.5-35B-A3B medium | Qwen | 1 | 0.6 | $0.716 | 11/23 | 111.5s |