智能体任务：答案错误
智能体任务
答案错误
看看哪些 AI 模型在 智能体任务 上最容易遇到 答案错误，更快找出薄弱点。
108/108
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） |
|---|---|---|---|---|---|---|---|
| #142 | Claude Opus 4.8 low | Anthropic | 1 | 4.7 | $3.787 | 0/1 | 110.7s |
| #143 | Qwen3.7 Flash high | Qwen | 1 | 4.7 | $0.063 | 0/1 | 135.9s |
| #144 | LongCat 2.0 medium | Meituan | 1 | 6.1 | $0.593 | 0/1 | 174.1s |
| #145 | GLM 5 medium | Z.ai | 1 | 4.7 | $0.360 | 0/1 | 110.7s |
| #149 | Claude Opus 5 none | Anthropic | 1 | 5.4 | $2.766 | 0/1 | 79.6s |
| #151 | MiMo-V2.6-Flash low | Xiaomi | 1 | 4.7 | $0.200 | 0/1 | 107.9s |
| #155 | Gemini 3.1 Flash Lite Preview medium | 1 | 5.0 | $0.158 | 0/1 | 54.9s | |
| #156 | Kimi K2.7 Code medium | Moonshot AI | 1 | 5.0 | $0.861 | 0/1 | 104.0s |
| #159 | Grok 4.6 low | X AI | 1 | 5.0 | $0.781 | 0/1 | 58.0s |
| #160 | Gemini 3 Flash Preview low | 1 | 5.0 | $0.254 | 0/1 | 52.6s | |
| #162 | Claude Opus 4.8 none | Anthropic | 1 | 5.4 | $2.328 | 0/1 | 69.1s |
| #163 | Gemini 3.1 Flash Lite medium | 1 | 5.0 | $0.153 | 0/1 | 47.8s | |
| #168 | Kimi K2.6 medium | Moonshot AI | 1 | 5.0 | $1.323 | 0/1 | 233.9s |
| #172 | GPT-6 Sol none | OpenAI | 1 | 5.0 | $0.551 | 0/1 | 57.0s |
| #173 | Space Bunny Alpha max | Stealth | 1 | 6.1 | $0.000 | 0/1 | 90.6s |