无效工具调用 失败
看看哪些 AI 模型最常遇到 无效工具调用，让你在选择前先发现稳定性风险。 排序方式: 总成本 ↑.
190/190
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 | 分数 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #90 | GPT-5.6 Terra high | OpenAI | 1 | 0.8 | $0.698 | 15/23 | 13.6s |
| #245 | Qwen3.5-35B-A3B medium | Qwen | 1 | 0.6 | $0.716 | 11/23 | 111.5s |
| #249 | Gemini 3.1 Flash Lite low | 1 | 0.6 | $0.717 | 13/23 | 19.0s | |
| #248 | Gemini 3.1 Flash Lite Preview low | 1 | 0.6 | $0.736 | 14/23 | 19.0s | |
| #273 | KAT-Coder-Pro V2.5 default | Kwaipilot | 1 | 0.6 | $0.742 | 11/23 | 25.0s |
| #45 | Gemini 3.5 Flash low | 1 | 0.9 | $0.752 | 19/23 | 7.85s | |
| #111 | Gemini 3.8 Flash medium | 1 | 0.8 | $0.769 | 19/23 | 16.6s | |
| #149 | Qwen3.6 Flash medium | Qwen | 1 | 0.7 | $0.783 | 13/23 | 45.5s |
| #292 | Trinity Large Thinking high | Arcee AI | 2 | 0.6 | $0.794 | 6/23 | 74.5s |
| #223 | Qwen3.6 35B A3B medium | Qwen | 1 | 0.7 | $0.834 | 12/23 | 59.9s |
| #237 | Kimi K2.5 medium | Moonshot AI | 1 | 0.6 | $0.841 | 10/23 | 125.8s |
| #267 | Trinity Large Thinking medium | Arcee AI | 2 | 0.6 | $0.862 | 8/23 | 84.0s |
| #71 | Muse Spark 1.1 low | Meta | 1 | 0.9 | $0.885 | 14/23 | 13.2s |
| #194 | Claude Sonnet 5.5 low | Anthropic | 1 | 0.7 | $0.909 | 10/23 | 7.46s |
| #77 | Muse Spark 1.3 low | Meta | 1 | 0.9 | $0.911 | 16/23 | 17.7s |