.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.

← All models
Tencent · 6 providers · 11.3M tool calls a day
#ProviderBroken
tool calls
Tool calls
a day
$ in / out
per M
Tokens/sUptime
1DeepInfrafp4#0.10%90k$0.79 / $2.383999.9%
2NovitaAIfp8#0.15%238k$0.83 / $2.5042100.0%
3Tencent Cloudfp8#0.16%10.7M$0.75 / $2.254699.9%
4Cohere#0.17%39k$0.83 / $2.507699.0%
5AtlasCloudfp8#0.35%13k$0.83 / $2.503799.9%
6SiliconFlowfp8#0.77%165k$0.83 / $2.503299.9%
Recommended OpenRouter settings for Hy4 preview
{
  "provider": {
    "order": [
      "deepinfra/fp4",
      "novita/fp8",
      "tencent/fp8"
    ],
    "ignore": []
  }
}

Point your agent at llms.txt

Using Hy4 preview?

Get updates when its successor drops

And which provider to use.

For model makers

Broken tool calls carry your model's name

Releasing a model? Test it before launch.

Worst provider0.77%

of Hy4 preview's tool calls fail on SiliconFlow (fp8).

Best to worst8×

as often on SiliconFlow (fp8) as on DeepInfra (fp4), its best provider.

Every day19k

broken tool calls a day across its 6 providers.