AI BENCHY 失败分析
答案错误 失败
看看哪些 AI 模型最常遇到 答案错误，让你在选择前先发现稳定性风险。 排序方式: 分数 ↑.
| 排名 | 模型 | 公司 | 答案错误 次数 | 分数 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|
| #73 | Seed-2.0-Mini medium | Bytedance Seed | 4 | 6.9 | 11/21 | 80.2s |
| #72 | DeepSeek V3.2 medium | DeepSeek | 5 | 7.0 | 11/21 | 68.7s |
| #70 | GPT-5.4 Nano medium | OpenAI | 8 | 7.0 | 11/21 | 12.0s |
| #71 | Step 3.7 Flash high | Stepfun | 6 | 7.0 | 11/21 | 64.5s |
| #68 | Claude Opus 4.8 none | Anthropic | 4 | 7.0 | 12/21 | 3.47s |
| #69 | Claude Opus 4.6 medium | Anthropic | 3 | 7.0 | 12/21 | 25.9s |
| #67 | MiniMax M3 medium | Minimax | 3 | 7.1 | 11/21 | 68.2s |
| #66 | Qwen3.5-35B-A3B medium | Qwen | 2 | 7.1 | 11/21 | 72.6s |
| #65 | Grok 4.20 medium | X AI | 6 | 7.1 | 12/21 | 27.7s |
| #64 | MiMo-V2-Flash medium | Xiaomi | 5 | 7.2 | 12/21 | 20.1s |
| #63 | GPT-5.3 Chat none | OpenAI | 7 | 7.2 | 12/21 | 6.34s |
| #62 | Step 3.5 Flash medium | Stepfun | 4 | 7.2 | 11/20 | 72.5s |
| #61 | Gemini 3.1 Flash Lite low | 9 | 7.2 | 12/21 | 1.89s | |
| #60 | Kimi K2.6 medium | Moonshot AI | 3 | 7.2 | 12/21 | 71.7s |
| #59 | GLM 5V Turbo medium | Z.ai | 7 | 7.2 | 11/21 | 23.1s |