指令遵循 排名
看看哪些 AI 模型在 指令遵循 上表现最好，哪些更稳定，以及差距主要出现在哪里。 排序方式: 响应时间（平均） ↓.
455/455
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 | 指令遵循 得分 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #8 | GPT-6 Astra xhigh | OpenAI | 1.0 | 1.0 | $6.213 | 2/2 | 4.91s |
| #358 | Grok 4.20 Beta medium | X AI | 1.0 | 0.5 | $0.750 | 2/2 | 4.89s |
| #116 | Gemini 3.8 Flash medium | 1.0 | 0.8 | $0.769 | 2/2 | 4.81s | |
| #27 | Claude Opus 5 high | Anthropic | 1.0 | 0.9 | $4.524 | 2/2 | 4.78s |
| #331 | Step 3.5 Flash medium | Stepfun | 0.8 | 0.5 | $0.105 | 1/2 | 4.78s |
| #2 | Gemini 3.8 Flash high | 1.0 | 1.0 | $1.927 | 2/2 | 4.73s | |
| #405 | Grok 4.1 Fast medium | X AI | 0.7 | 0.4 | $0.069 | 1/2 | 4.63s |
| #75 | Claude Sonnet 5.5 xhigh | Anthropic | 1.0 | 0.9 | $1.516 | 2/2 | 4.58s |
| #91 | DeepSeek V4 Flash 0731 high | DeepSeek | 1.0 | 0.8 | $0.576 | 2/2 | 4.46s |
| #56 | Muse Spark 1.2 medium | Meta | 1.0 | 0.9 | $1.590 | 2/2 | 4.40s |
| #140 | Nemotron 3.5 Super VL medium | NVIDIA | 1.0 | 0.8 | $0.437 | 2/2 | 4.34s |
| #78 | Qwen3.8 Max (0902) low | Qwen | 1.0 | 0.9 | $1.131 | 2/2 | 4.32s |
| #128 | Ling 3.1 Flash high | Inclusionai | 1.0 | 0.8 | $0.000 | 2/2 | 4.31s |
| #406 | Laguna M.1 medium | Poolside | 1.0 | 0.4 | $0.033 | 2/2 | 4.30s |
| #345 | MiMo-V2-Flash medium | Xiaomi | 1.0 | 0.5 | $0.043 | 2/2 | 4.28s |