of Qwen3.5-9B's tool calls fail on Prime Intellect (bf16).
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 | Gemma 4 26B A4Bbf16 | 0.02% | 3 of 11 | NextBit (bf16) 0.00% | 55k |
| 2 | DeepSeek V4.1 Flash 1017 | 0.04% | 2 of 35 | Sail Research (fp4) 0.03% | 311k |
| 3 | DeepSeek V4.1 Flash | 0.06% | 1 of 35 | — | 139k |
| 4 | GLM 5.3 | 0.38% | 12 of 42 | Amazon Bedrock 0.08% | 57k |
| 5 | GLM 5.3 Flash | 0.49% | 12 of 34 | NextBit (fp4) 0.00% | 171k |
| 6 | Qwen3.5-9Bbf16 | 0.58% | 3 of 5 | Darkbloom (fp4) 0.08% | 86k |
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 Gemma 4 26B A4B, its best model.
models, where Prime Intellect is the most reliable provider.