AI BENCHY 失败分析
API 错误 失败
看看哪些 AI 模型最常遇到 API 错误，让你在选择前先发现稳定性风险。 排序方式: 测试正确 ↑.
| 排名 | 模型 | 公司 | API 错误 次数 | 分数 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|
| #99 | Step 3.5 Flash none | Stepfun | 1 | 3.0 | 0/1 | 0ms |
| #98 | LFM2-24B-A2B none | Liquid | 4 | 4.1 | 1/16 | 811ms |
| #94 | MiMo-V2-Flash none | Xiaomi | 1 | 4.5 | 3/18 | 2.79s |
| #84 | gpt-oss-120b none | OpenAI | 3 | 5.2 | 4/18 | 12.0s |
| #73 | Mistral Small 4 medium | Mistral | 2 | 5.7 | 5/18 | 5.64s |
| #72 | Hunter Alpha none | OpenRouter | 1 | 5.7 | 6/18 | 4.58s |
| #56 | Grok 4.20 Multi Agent Beta medium | X AI | 2 | 6.4 | 7/18 | 9.80s |
| #50 | Hunter Alpha medium | OpenRouter | 1 | 6.7 | 8/18 | 10.3s |
| #47 | Grok 4.20 medium | X AI | 1 | 7.0 | 9/18 | 10.3s |
| #51 | Nemotron 3 Super medium | NVIDIA | 1 | 6.7 | 9/18 | 19.1s |
| #43 | Qwen3.5-35B-A3B medium | Qwen | 1 | 7.4 | 10/18 | 44.5s |
| #48 | Gemma 4 31B none | 2 | 6.9 | 10/18 | 4.02s | |
| #32 | Qwen3.5-Flash medium | Qwen | 1 | 7.8 | 11/18 | 66.7s |
| #41 | MiMo-V2-Flash medium | Xiaomi | 1 | 7.5 | 11/18 | 23.4s |
| #33 | GLM 5.1 medium | Z.ai | 1 | 7.8 | 12/18 | 24.1s |