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: 分数 ↓.
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） |
|---|---|---|---|---|---|---|---|
| #155 | DeepSeek V4 Pro 0423 none | DeepSeek | 1 | 0.7 | $0.083 | 10/23 | 14.1s |
| #156 | Inkling medium | Thinkingmachines | 2 | 0.7 | $0.527 | 15/23 | 17.0s |
| #157 | Gemma 4 26B A4B medium | 1 | 0.7 | $0.101 | 16/23 | 105.0s | |
| #158 | GLM 5.3 Flash high | Z.ai | 1 | 0.7 | $0.056 | 13/23 | 29.1s |
| #164 | Gemini 3.5 Flash Lite medium | 2 | 0.7 | $0.440 | 14/23 | 8.68s | |
| #166 | Qwen3.5 Plus 2026-02-15 medium | Qwen | 1 | 0.7 | $0.567 | 14/23 | 97.7s |
| #167 | Qwen3.8 27B medium | Qwen | 1 | 0.7 | ~$0.073 | 14/23 | 40.1s |
| #170 | Qwen3.7 Flash high | Qwen | 1 | 0.7 | $0.063 | 13/23 | 45.5s |
| #173 | Step 3.7 Flash low | Stepfun | 2 | 0.7 | $0.449 | 11/23 | 20.7s |
| #174 | Qwen3.6 27B medium | Qwen | 2 | 0.7 | $1.021 | 11/23 | 107.8s |
| #175 | Claude Opus 5 none | Anthropic | 1 | 0.7 | $2.766 | 12/23 | 10.8s |
| #176 | Qwen3.8 Flash Next xhigh | Qwen | 1 | 0.7 | ~$0.195 | 14/23 | 102.7s |
| #177 | GLM 5.3 FlashX medium | Z.ai | 1 | 0.7 | $0.114 | 12/23 | 7.37s |
| #178 | MiMo-V2.6-Flash low | Xiaomi | 1 | 0.7 | $0.200 | 12/23 | 54.6s |
| #179 | MiMo-V2.5 medium | Xiaomi | 1 | 0.7 | $0.130 | 12/23 | 49.2s |