of gpt-oss-20b's tool calls fail on Parasail (fp4).
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.
| # | Provider | Broken tool calls | Tool calls a day | $ in / out per M | Tokens/s | Uptime |
|---|---|---|---|---|---|---|
| 1 | Darkbloomfp8# | 0.13% | 20k | $0.02 / $0.09 | 62 | 100.0% |
| 2 | Groq# | 0.33% | 307k | $0.07 / $0.30 | 394 | 99.8% |
| 3 | DeepInfrabf16# | 2.01% | 5.3k | $0.03 / $0.14 | 64 | 99.3% |
| 4 | Amazon Bedrock# | 2.77% | 191k | $0.07 / $0.15 | 320 | 100.0% |
| 5 | CoreWeavefp4# | 13.45% | 37k | $0.03 / $0.13 | 96 | 99.9% |
| 6 | AkashMLfp4# | 20.78% | 6.2k | $0.02 / $0.10 | 136 | 100.0% |
| 7 | Parasailfp4# | 22.66% | 57k | $0.03 / $0.15 | 119 | 96.8% |
Recommended OpenRouter settings for gpt-oss-20b
{
"provider": {
"order": [
"darkbloom/fp8",
"groq",
"deepinfra/bf16"
],
"ignore": [
"coreweave/fp4",
"akashml/fp4",
"parasail/fp4"
]
}
}Point your agent at llms.txt
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.
as often on Parasail (fp4) as on Darkbloom (fp8), its best provider.
broken tool calls a day across its 7 providers.