of MiMo-V2.6-Flash's tool calls fail on Makora (fp8).
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 K3fp4 | 0.32% | 19 of 24 | DeepInfra (mxfp4) 0.01% | 26k |
| 2 | DeepSeek V4.1 Flashfp8 | 0.70% | 20 of 35 | Prime Intellect 0.06% | 43k |
| 3 | GLM 5.3fp4 | 0.79% | 23 of 42 | Amazon Bedrock 0.08% | 32k |
| 4 | DeepSeek V4.1 Flash 1017fp8 | 0.97% | 22 of 35 | Sail Research (fp4) 0.03% | 202k |
| 5 | GLM 5.3 Flash | 1.25% | 19 of 34 | NextBit (fp4) 0.00% | 275k |
| 6 | Muse Spark Openfp4 | 1.61% | 4 of 6 | Nebius Token Factory (fp4) 0.07% | 38k |
| 7 | Gemma 4 26B A4B | 2.48% | 7 of 11 | NextBit (bf16) 0.00% | 29k |
| 8 | MiMo-V2.6-Flashfp8 | 3.92% | 7 of 9 | Venice (fp8) 1.41% | 149k |
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 Makora.