指令遵循：API 错误
指令遵循
API 错误
看看哪些 AI 模型在 指令遵循 上最容易遇到 API 错误，更快找出薄弱点。
21/21
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 | API 错误 次数 | 分类得分 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #408 | Command A+ medium | Cohere | 2 | 0.3 | $0.064 | 0/2 | 103.0s |
| #414 | Command A+ high | Cohere | 2 | 0.3 | $0.079 | 0/2 | 110.3s |
| #417 | Command A+ none | Cohere | 2 | 0.3 | $0.000 | 0/2 | 95.7s |
| #418 | Command A+ low | Cohere | 2 | 0.3 | $0.073 | 0/2 | 105.7s |
| #32 | Muse Wonder medium | Meta | 1 | 0.7 | $3.397 | 1/2 | 3.88s |
| #40 | Muse Wonder low | Meta | 1 | 0.7 | $2.637 | 1/2 | 2.08s |
| #43 | Muse Wonder minimal | Meta | 1 | 0.7 | $1.639 | 1/2 | 1.39s |
| #131 | Muse Wonder high | Meta | 1 | 0.7 | $3.426 | 1/2 | 1.97s |
| #139 | Muse Wonder xhigh | Meta | 1 | 0.7 | $4.012 | 1/2 | 2.64s |
| #141 | GPT-6 Luna low | OpenAI | 1 | 0.7 | $0.033 | 1/2 | 17.4s |
| #184 | Ember-1 low | Fireworks | 1 | 0.6 | $1.488 | 1/2 | 32.7s |
| #185 | Step 5 Preview high | Stepfun | 1 | 0.7 | $4.513 | 1/2 | 20.5s |
| #191 | Ember-1 high | Fireworks | 1 | 0.7 | $2.405 | 1/2 | 33.4s |
| #209 | Ember-1 max | Fireworks | 1 | 0.6 | $3.264 | 1/2 | 33.0s |
| #225 | Mercury 2.5 high | Inception | 1 | 0.7 | $0.044 | 1/2 | 934ms |