of Qwen3.5 397B A17B's tool calls fail on Phala.
Updated October 29, 2026
Tool-call index
Not all models are served equally by all providers. On a typical model, the worst provider breaks tool calls 19× as often as the best. Find the best one for your agents.
| # | Model | Broken tool calls | Rank | Best on the model | Tool calls a day |
|---|---|---|---|---|---|
| 1 | Kimi K3 | 0.17% | 14 of 24 | DeepInfra (mxfp4) 0.01% | 129k |
| 2 | DeepSeek V4 Pro 0813 | 0.75% | 9 of 21 | Ionstream 0.04% | 2.2k |
| 3 | DeepSeek V3.2 | 0.76% | 3 of 4 | SiliconFlow (fp8) 0.60% | 3.0k |
| 4 | Qwen3.8 27B | 0.86% | 9 of 21 | Darkbloom (fp4) 0.07% | 47k |
| 5 | GLM 5.3 Flashnvfp4 | 0.89% | 15 of 34 | NextBit (fp4) 0.00% | 612k |
| 6 | Hy3 | 1.71% | 4 of 5 | Tencent Cloud (fp8) 1.46% | 5.9k |
| 7 | DeepSeek V4.1 Flash 1017 | 1.92% | 29 of 35 | Sail Research (fp4) 0.03% | 64k |
| 8 | DeepSeek V4.1 Flash | 2.74% | 33 of 35 | Prime Intellect 0.06% | 32k |
| 9 | GLM 5.3 | 4.43% | 40 of 42 | Amazon Bedrock 0.08% | 7.3k |
| 10 | Qwen3.6 35B A3B | 5.75% | 6 of 6 | AkashML (fp8) 1.63% | 8.0k |
| 11 | DeepSeek V4 Flash 0731 | 7.10% | 20 of 22 | Sail Research (fp4) 0.04% | 9.8k |
| 12 | Qwen3.5 397B A17B | 11.30% | 7 of 7 | Alibaba Cloud Int. 0.58% | 6.1k |
Get updates when a new model drops
And which provider to use.
For inference providers
Tool calling doesn't have to break
The same provider can handle tool calls well on one model and struggle on the next.
on Kimi K3, its best model.
models, where the settings tell OpenRouter to skip Phala.