.TXT Serve models?↓Release models?↓
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.

openrouter-data-staging.dottxt.workers.dev/llms.txt

Models

Error rate on the best and worst provider.
New models

Get updates when a new model drops

And which provider to use.

Inference providers

Error rate on the best and worst model.
For inference providers

Tool calling doesn't have to break

Worst provider89.86%

of Qwen3.5-9B's tool calls fail on its worst provider.

Over 3%34 of 57

providers break 3% of tool calls or more on at least one model.

Same provider14×

between a typical provider's error rate on its best model and on its worst.

Fix this with dottxt↵
For model makers

Broken tool calls carry your model's name

The worst error rate on each maker's models.

DeepSeek15.01%

of DeepSeek V4 Flash 0731's tool calls fail on Alibaba Cloud Int.

Google8.70%

of Gemma 4 26B A4B's tool calls fail on AkashML (fp4).

IBM3.71%

of Granite 4.2 8B's tool calls fail on CoreWeave (bf16).

Meta67.14%

of Llama 3.1 8B Instruct's tool calls fail on CoreWeave (bf16).

MiniMax36.08%

of MiniMax M3's tool calls fail on AtlasCloud (fp8).

Mistral1.38%

of Mistral Nemo's tool calls fail on io.net (fp16).

MoonshotAI6.65%

of Kimi K2.5's tool calls fail on Amazon Bedrock.

NVIDIA9.96%

of Nemotron 3.5 Lightning's tool calls fail on DeepInfra (bf16).

OpenAI80.12%

of gpt-oss-120b's tool calls fail on Nebius Token Factory (fp4).

Qwen89.86%

of Qwen3.5-9B's tool calls fail on SiliconFlow (fp8).

Reflection6.52%

of Beam 1.0 501B A23B's tool calls fail on CoreWeave (fp4).

Tencent3.23%

of Hy3's tool calls fail on NovitaAI.

Xiaomi5.47%

of MiMo-V2.6-Pro's tool calls fail on NovitaAI (fp8).

Z.ai23.02%

of GLM 5.3's tool calls fail on GMICloud (fp8).

Releasing a model? Test it before launch.

Work with us on your release↵