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 ↑.
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） |
|---|---|---|---|---|---|---|---|
| #58 | Gemini 2.5 Flash medium | 1 | 8.2 | $0.644 | 15/22 | 21.1s | |
| #57 | Claude Fable 5.1 medium | Anthropic | 1 | 8.3 | $2.909 | 15/22 | 11.7s |
| #56 | Muse Spark 1.1 low | Meta | 3 | 8.3 | $0.673 | 13/22 | 11.9s |
| #55 | Inkling Small high | Thinkingmachines | 2 | 8.4 | $0.398 | 14/22 | 30.4s |
| #54 | Seed 2.1 Turbo medium | Bytedance Seed | 1 | 8.4 | $1.951 | 15/22 | 193.6s |
| #53 | Qwen3.8 27B high | Qwen | 2 | 8.4 | ~$0.287 | 16/22 | 167.9s |
| #52 | GPT-5.2 medium | OpenAI | 3 | 8.4 | $1.182 | 14/22 | 28.5s |
| #50 | Grok 4.6 medium | X AI | 1 | 8.4 | $1.713 | 16/22 | 73.5s |
| #49 | Muse Spark 1.2 low | Meta | 3 | 8.4 | $0.655 | 15/22 | 9.77s |
| #47 | GPT-5.4 medium | OpenAI | 2 | 8.5 | $1.560 | 15/22 | 22.9s |
| #46 | Qwen3.8 Max low | Qwen | 1 | 8.6 | $0.702 | 16/22 | 21.8s |
| #43 | Muse Spark 1.1 medium | Meta | 2 | 8.6 | $1.420 | 15/22 | 26.2s |
| #42 | Muse Spark 1.2 medium | Meta | 3 | 8.6 | $1.339 | 14/22 | 19.7s |
| #41 | Muse Spark 1.2 high | Meta | 3 | 8.7 | $1.771 | 14/22 | 26.6s |
| #39 | GLM 5.3 Flash max | Z.ai | 1 | 8.7 | $0.059 | 16/22 | 52.1s |