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 ↑.
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） |
|---|---|---|---|---|---|---|---|
| #210 | Qwen3.5-9B medium | Qwen | 1 | 3.8 | $0.036 | 3/22 | 82.2s |
| #211 | Laguna Xs.2 none | Poolside | 1 | 3.8 | $0.004 | 5/19 | 806ms |
| #213 | Nemotron 3 Nano Omni 30b A3b Reasoning medium | NVIDIA | 1 | 3.4 | $0.000 | 4/19 | 17.1s |
| #215 | Step 3.5 Flash none | Stepfun | 1 | 2.3 | $0.020 | 6/12 | 39.0s |
| #216 | LFM2-24B-A2B none | Liquid | 1 | 2.2 | $0.001 | 2/16 | 782ms |
| #16 | GPT-5.3-Codex medium | OpenAI | 2 | 8.9 | $0.920 | 16/22 | 17.0s |
| #19 | Muse Spark 1.1 medium | Meta | 2 | 8.6 | $1.357 | 15/22 | 25.0s |
| #21 | GPT-5.4 medium | OpenAI | 2 | 8.5 | $1.533 | 15/22 | 23.1s |
| #27 | Muse Spark 1.1 low | Meta | 2 | 8.3 | $0.647 | 13/22 | 11.5s |
| #30 | Muse Spark 1.1 high | Meta | 2 | 8.1 | $1.694 | 12/22 | 31.5s |
| #39 | Seed-2.0-Lite medium | Bytedance Seed | 2 | 7.9 | $0.234 | 14/22 | 48.5s |
| #49 | DeepSeek V4 Flash high | DeepSeek | 2 | 7.7 | $0.041 | 13/22 | 49.7s |
| #50 | DeepSeek V4 Pro high | DeepSeek | 2 | 7.7 | $0.200 | 10/22 | 79.1s |
| #51 | MiniMax M3 medium | Minimax | 2 | 7.6 | $0.286 | 12/22 | 75.0s |
| #57 | GPT-5.4 Nano medium | OpenAI | 2 | 7.5 | $0.138 | 12/22 | 13.2s |