谜题求解：答案错误
谜题求解
答案错误
看看哪些 AI 模型在 谜题求解 上最容易遇到 答案错误，更快找出薄弱点。 排序方式: 失败次数 ↑.
142/142
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 | 分类得分 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #3 | GPT-5.6 Sol low | OpenAI | 1 | 8.2 | $0.971 | 2/3 | 3.44s |
| #4 | GPT-5.6 Sol medium | OpenAI | 1 | 8.2 | $1.316 | 2/3 | 2.98s |
| #9 | Gemini 3.5 Flash medium | 1 | 7.7 | $0.642 | 2/3 | 2.38s | |
| #23 | Claude Sonnet 5 medium | Anthropic | 1 | 7.7 | $0.922 | 2/3 | 2.98s |
| #25 | Gemini 2.5 Flash medium | 1 | 7.7 | $0.643 | 2/3 | 3.18s | |
| #26 | GPT-5 Mini medium | OpenAI | 1 | 5.6 | $0.237 | 1/3 | 15.2s |
| #28 | Inkling high | Thinkingmachines | 1 | 6.9 | $1.006 | 1/3 | 10.7s |
| #30 | GPT-5.2 Chat none | OpenAI | 1 | 7.7 | $0.604 | 2/3 | 4.10s |
| #31 | GLM 5.2 high | Z.ai | 1 | 6.0 | $0.970 | 1/3 | 33.7s |
| #34 | GPT-5.6 Terra high | OpenAI | 1 | 7.7 | $1.055 | 2/3 | 5.45s |
| #38 | GLM 5.2 medium | Z.ai | 1 | 8.2 | $0.222 | 2/3 | 13.1s |
| #39 | GPT-5.6 Terra medium | OpenAI | 1 | 8.4 | $0.676 | 2/3 | 3.78s |
| #43 | Claude Opus 4.6 medium | Anthropic | 1 | 7.7 | $3.059 | 2/3 | 4.71s |
| #44 | GPT-5.6 Luna high | OpenAI | 1 | 7.6 | $1.017 | 2/3 | 14.6s |
| #46 | DeepSeek V4 Pro high | DeepSeek | 1 | 6.9 | $0.200 | 1/3 | 56.8s |