of Qwen3 235B A22B Instruct 2507's tool calls fail on Venice (fp8).
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.
| # | Provider | Broken tool calls | Tool calls a day | $ in / out per M | Tokens/s | Uptime |
|---|---|---|---|---|---|---|
| 1 | Parasailfp8# | 0.78% | 77k | $0.14 / $0.80 | 40 | 100.0% |
| 2 | DeepInfrafp8# | 1.09% | 76k | $0.09 / $0.55 | 7 | 95.2% |
| 3 | Venicefp8# | 2.21% | 6.8k | $0.15 / $0.75 | 34 | 97.1% |
{
"provider": {
"order": [
"parasail/fp8",
"deepinfra/fp8",
"venice/fp8"
],
"ignore": []
}
}Point your agent at llms.txt
And which provider to use.
Releasing a model? Test it before launch.
of Qwen3 235B A22B Instruct 2507's tool calls fail on Venice (fp8).
as often on Venice (fp8) as on Parasail (fp8), its best provider.
broken tool calls a day across its 3 providers.