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

← All models

Meta · 5 providers · 74k tool calls a day

#	Provider	Broken tool calls	Tool calls a day	$ in / out per M	Tokens/s	Uptime
1	Groq#	2.21%	2.5k	$0.59 / $0.79	223	100.0%
2	CoreWeavefp16#	4.65%	54k	$0.71 / $0.71	87	100.0%
3	DeepInfrafp8/turbo#	9.94%	6.5k	$0.10 / $0.32	19	99.2%
4	NovitaAIbf16#	25.58%	2.4k	$0.14 / $0.40	35	98.8%
5	AkashMLfp8#	32.75%	9.0k	$0.20 / $0.52	33	100.0%

Recommended OpenRouter settings for Llama 3.3 70B Instruct

{
  "provider": {
    "order": [
      "groq"
    ],
    "ignore": [
      "coreweave/fp16",
      "deepinfra/turbo",
      "novita/bf16",
      "akashml/fp8"
    ]
  }
}

Point your agent at llms.txt

Using Llama 3.3 70B Instruct?

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 provider32.75%

of Llama 3.3 70B Instruct's tool calls fail on AkashML (fp8).

Best to worst15×

as often on AkashML (fp8) as on Groq, its best provider.

Every day7k

broken tool calls a day across its 5 providers.

Work with us on your release↵