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: 测试正确 ↓.
216/216
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） |
|---|---|---|---|---|---|---|---|
| #127 | gpt-oss-120b medium | OpenAI | 9.9 | 6.1 | $0.019 | 2/2 | 7.63s |
| #128 | Gemini 3.1 Flash Lite none | 10.0 | 6.1 | $0.046 | 2/2 | 859ms | |
| #129 | Inkling low | Thinkingmachines | 9.8 | 6.1 | $0.187 | 2/2 | 1.81s |
| #134 | GPT-5 Nano medium | OpenAI | 9.8 | 6.1 | $0.114 | 2/2 | 15.6s |
| #135 | Nemotron 3 Ultra none | NVIDIA | 10.0 | 6.1 | $0.095 | 2/2 | 1.46s |
| #137 | Grok 4.20 Beta medium | X AI | 9.8 | 6.0 | $0.750 | 2/2 | 4.89s |
| #139 | Gemini 3 PRO Preview medium | 9.8 | 6.0 | $0.385 | 2/2 | 3.26s | |
| #141 | Hy3 preview high | Tencent | 10.0 | 5.9 | $0.048 | 2/2 | 34.4s |
| #143 | North Mini Code medium | Cohere | 9.8 | 5.9 | $0.000 | 2/2 | 15.4s |
| #147 | GLM 5 none | Z.ai | 10.0 | 5.7 | $0.041 | 2/2 | 1.48s |
| #150 | KAT-Coder-Air V2.5 high | Kwaipilot | 9.8 | 5.6 | $0.077 | 2/2 | 1.51s |
| #155 | KAT-Coder-Air V2.5 medium | Kwaipilot | 10.0 | 5.6 | $0.048 | 2/2 | 1.50s |
| #157 | GLM 5.1 none | Z.ai | 9.8 | 5.5 | $0.164 | 2/2 | 1.98s |
| #159 | Hy3 preview low | Tencent | 10.0 | 5.5 | $0.015 | 2/2 | 16.0s |
| #164 | KAT-Coder-Air V2.5 low | Kwaipilot | 9.8 | 5.4 | $0.041 | 2/2 | 1.64s |