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: 分数 ↓.
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） |
|---|---|---|---|---|---|---|---|
| #279 | Step 5 Preview medium | Stepfun | 2 | 0.6 | $4.136 | 10/23 | 136.1s |
| #280 | Granite 4.2 8B medium | IBM Granite | 1 | 0.6 | $0.029 | 8/23 | 50.5s |
| #282 | Apodex 1.1 Mini medium | Apodex | 2 | 0.6 | $0.000 | 11/23 | 32.7s |
| #283 | Claude Sonnet 4.6 none | Anthropic | 1 | 0.6 | $0.661 | 12/22 | 7.67s |
| #285 | MiMo-V2.6-Pro none | Xiaomi | 1 | 0.6 | $0.193 | 9/23 | 9.66s |
| #286 | Inkling Small medium | Thinkingmachines | 4 | 0.6 | $0.113 | 10/23 | 6.03s |
| #288 | Gemini 3.1 Flash Lite minimal | 3 | 0.6 | $0.080 | 10/23 | 3.95s | |
| #289 | gpt-oss-120b medium | OpenAI | 3 | 0.6 | $0.030 | 9/23 | 31.6s |
| #290 | Gemini 3.1 Flash Lite none | 1 | 0.6 | $0.069 | 9/23 | 3.58s | |
| #291 | Qwen3.6 Flash none | Qwen | 1 | 0.6 | $0.139 | 7/23 | 7.63s |
| #292 | Claude Sonnet 5 none | Anthropic | 1 | 0.6 | $1.095 | 7/23 | 8.60s |
| #293 | Nemotron 3 Ultra none | NVIDIA | 1 | 0.6 | $0.164 | 8/23 | 6.76s |
| #294 | Qwen3.7 Flash none | Qwen | 1 | 0.6 | $0.028 | 7/23 | 14.7s |
| #296 | DeepSeek V4.1 Flash 1017 none | DeepSeek | 1 | 0.6 | $0.055 | 8/23 | 3.46s |
| #297 | GPT-5.6 Terra none | OpenAI | 1 | 0.6 | $0.377 | 8/23 | 4.11s |