智能体任务：无效工具调用
智能体任务
无效工具调用
看看哪些 AI 模型在 智能体任务 上最容易遇到 无效工具调用，更快找出薄弱点。
44/44
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 | 无效工具调用 次数 | 分类得分 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #46 | Claude Opus 5 medium | Anthropic | 1 | 7.4 | $2.896 | 0/1 | 87.6s |
| #52 | Gemini 3.7 Flash low | 1 | 6.4 | $0.293 | 0/1 | 62.6s | |
| #61 | Qwen3.8 Max (0902) low | Qwen | 1 | 6.1 | $1.131 | 0/1 | 136.0s |
| #67 | Gemini 3.5 Flash high | 1 | 4.1 | $2.672 | 0/1 | 84.7s | |
| #85 | Claude Sonnet 5.5 high | Anthropic | 1 | 7.4 | $1.197 | 0/1 | 53.6s |
| #87 | Claude Opus 4.7 medium | Anthropic | 1 | 5.2 | $3.636 | 0/1 | 150.6s |
| #89 | Gemini 3.5 Flash medium | 1 | 3.8 | $1.308 | 0/1 | 83.6s | |
| #97 | GLM 5.3 FlashX max | Z.ai | 1 | 6.1 | $0.623 | 0/1 | 104.8s |
| #106 | GPT-6 Luna medium | OpenAI | 1 | 7.4 | $0.048 | 0/1 | 71.7s |
| #110 | GPT-6 Luna high | OpenAI | 1 | 5.2 | $0.074 | 0/1 | 82.8s |
| #121 | GPT-5.4 Mini medium | OpenAI | 1 | 7.4 | $0.965 | 0/1 | 84.3s |
| #122 | GPT-5.6 Luna medium | OpenAI | 1 | 7.4 | $0.101 | 0/1 | 58.1s |
| #130 | Inkling medium | Thinkingmachines | 1 | 4.7 | $0.540 | 0/1 | 45.9s |
| #138 | Gemini 3.5 Flash Lite medium | 1 | 4.7 | $0.440 | 0/1 | 86.9s | |
| #146 | Step 3.7 Flash low | Stepfun | 1 | 7.4 | $0.449 | 0/1 | 83.6s |