of DeepSeek V3.1's tool calls fail on DeepInfra (fp4).
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 | SiliconFlowfp8# | 1.04% | 2.5k | $0.27 / $1.00 | 18 | 96.2% |
| 2 | CoreWeavefp8# | 2.27% | 2.0k | $0.55 / $1.65 | 82 | 100.0% |
| 3 | DeepInfrafp4# | 4.03% | 14k | $0.25 / $0.95 | 22 | 100.0% |
{
"provider": {
"order": [
"siliconflow/fp8",
"coreweave/fp8"
],
"ignore": [
"deepinfra/fp4"
]
}
}Point your agent at llms.txt
And which provider to use.
Releasing a model? Test it before launch.
of DeepSeek V3.1's tool calls fail on DeepInfra (fp4).
as often on DeepInfra (fp4) as on SiliconFlow (fp8), its best provider.
broken tool calls a day across its 3 providers.