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 ↑.
313/313
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） |
|---|---|---|---|---|---|---|---|
| #241 | GPT-5.6 Luna none | OpenAI | 14 | 5.4 | $0.029 | 6/22 | 1.50s |
| #248 | Seed-2.0-Code none | Bytedance Seed | 11 | 5.3 | $0.151 | 6/22 | 14.7s |
| #253 | Inkling none | Thinkingmachines | 13 | 5.2 | $0.147 | 6/22 | 3.47s |
| #265 | Mercury 2.5 Preview low | Inception | 12 | 5.0 | $0.011 | 6/22 | 1.31s |
| #267 | DeepSeek V3.2 none | DeepSeek | 8 | 5.0 | $0.054 | 6/22 | 17.9s |
| #271 | Ling-2.6-flash none | Inclusionai | 9 | 4.9 | $0.002 | 6/22 | 10.7s |
| #262 | GLM 5 Turbo none | Z.ai | 13 | 5.1 | $0.047 | 6/21 | 2.82s |
| #295 | Elephant Alpha medium | Openrouter | 9 | 4.3 | $0.000 | 6/21 | 1.27s |
| #299 | Laguna Xs.2 medium | Poolside | 6 | 4.1 | $0.015 | 6/19 | 6.73s |
| #312 | gpt-oss-120b none | OpenAI | 8 | 3.7 | $0.010 | 6/19 | 21.6s |
| #186 | Claude Sonnet 5 none | Anthropic | 8 | 6.1 | $0.548 | 7/22 | 6.05s |
| #190 | Qwen3.7 Flash none | Qwen | 13 | 6.1 | $0.019 | 7/22 | 10.1s |
| #195 | Qwen3.6 Flash none | Qwen | 12 | 6.1 | $0.062 | 7/22 | 3.73s |
| #205 | Qwen3.5-Flash none | Qwen | 14 | 6.0 | $0.073 | 7/22 | 25.3s |
| #214 | GPT-5.4 none | OpenAI | 14 | 5.8 | $0.397 | 7/22 | 2.08s |