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: 总成本 ↓.
141/141
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） |
|---|---|---|---|---|---|---|---|
| #196 | MiniMax M2.5 medium | Minimax | 3 | 4.6 | $0.340 | 5/22 | 68.3s |
| #115 | Mimo V2 PRO medium | Xiaomi | 1 | 6.3 | $0.333 | 12/21 | 22.2s |
| #53 | GLM 5 Turbo medium | Z.ai | 1 | 7.6 | $0.323 | 14/21 | 23.0s |
| #46 | GLM 5 medium | Z.ai | 1 | 7.7 | $0.307 | 15/21 | 33.5s |
| #92 | Gemini 3.5 Flash minimal | 1 | 6.8 | $0.300 | 14/22 | 2.65s | |
| #51 | MiniMax M3 medium | Minimax | 2 | 7.6 | $0.286 | 12/22 | 75.0s |
| #123 | GPT-5.6 Luna low | OpenAI | 1 | 6.2 | $0.249 | 10/22 | 5.04s |
| #148 | Qwen3.5-122B-A10B none | Qwen | 2 | 5.7 | $0.247 | 6/22 | 12.9s |
| #29 | GPT-5 Mini medium | OpenAI | 3 | 8.1 | $0.237 | 12/22 | 27.6s |
| #39 | Seed-2.0-Lite medium | Bytedance Seed | 2 | 7.9 | $0.234 | 14/22 | 48.5s |
| #50 | DeepSeek V4 Pro high | DeepSeek | 2 | 7.7 | $0.200 | 10/22 | 79.1s |
| #88 | MiMo-V2.5-Pro medium | Xiaomi | 2 | 6.9 | $0.187 | 12/22 | 33.9s |
| #129 | Inkling low | Thinkingmachines | 2 | 6.1 | $0.187 | 10/22 | 5.15s |
| #144 | Kimi K2.6 none | Moonshot AI | 3 | 5.8 | $0.184 | 7/22 | 19.6s |
| #200 | GLM 4.7 Flash medium | Z.ai | 2 | 4.3 | $0.166 | 4/22 | 142.6s |