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: 测试正确 ↓.
209/209
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 | 答案错误 次数 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #1 | Gemini 3 Flash Preview medium | 1 | 9.6 | $0.742 | 21/22 | 19.2s | |
| #2 | Gemini 3.5 Flash high | 1 | 9.5 | $1.976 | 20/22 | 15.1s | |
| #7 | Gemini 3.1 Pro Preview medium | 2 | 9.2 | $1.361 | 20/22 | 21.5s | |
| #6 | GPT-5.5 low | OpenAI | 3 | 9.3 | $1.253 | 19/22 | 10.1s |
| #9 | Gemini 3.5 Flash medium | 2 | 9.1 | $0.642 | 19/22 | 8.20s | |
| #11 | Gemini 3.5 Flash low | 2 | 8.9 | $0.433 | 19/22 | 5.55s | |
| #94 | Claude Opus 4.7 none | Anthropic | 3 | 6.6 | $0.505 | 16/19 | 3.02s |
| #3 | GPT-5.6 Sol low | OpenAI | 4 | 9.5 | $0.971 | 18/22 | 8.79s |
| #4 | GPT-5.6 Sol medium | OpenAI | 4 | 9.4 | $1.316 | 18/22 | 11.4s |
| #5 | GPT-5.6 Sol high | OpenAI | 4 | 9.4 | $1.234 | 18/22 | 11.7s |
| #8 | Qwen3.7 Max medium | Qwen | 3 | 9.2 | $1.116 | 18/22 | 40.6s |
| #10 | GPT-5.5 medium | OpenAI | 4 | 9.0 | $4.137 | 18/22 | 38.4s |
| #14 | Claude Opus 4.8 medium | Anthropic | 3 | 8.8 | $1.931 | 18/22 | 12.5s |
| #15 | Claude Opus 4.7 medium | Anthropic | 3 | 8.7 | $1.477 | 18/22 | 7.61s |
| #163 | Gemini 3.1 Flash Lite Preview high | 2 | 5.3 | $2.310 | 13/16 | 68.1s |