of DeepSeek V4 Pro 0423's tool calls fail on Baidu Qianfan (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 | Cloudflare# | 0.78% | 27k | $1.15 / $2.55 | 41 | 99.5% |
| 2 | SiliconFlowfp8# | 1.02% | 11k | $1.00 / $3.13 | 42 | 99.2% |
| 3 | Crusoefp8# | 1.18% | 4.6k | $1.74 / $3.48 | 58 | 97.0% |
| 4 | NovitaAIfp8# | 1.28% | 138k | $1.60 / $3.20 | 66 | 99.9% |
| 5 | GMICloudfp8# | 1.63% | 6.2k | $0.96 / $1.91 | 22 | 99.9% |
| 6 | Relacefp4# | 1.66% | 159k | $0.90 / $1.50 | 90 | 99.9% |
| 7 | DeepInfrafp8# | 1.69% | 13k | $1.30 / $2.60 | 66 | 100.0% |
| 8 | StreamLakefp8# | 1.75% | 326k | $0.82 / $1.64 | 44 | 100.0% |
| 9 | Alibaba Cloud Int.fp8# | 3.32% | 39k | $1.42 / $2.83 | 45 | 98.9% |
| 10 | AtlasCloudfp4# | 3.78% | 2.2k | $1.68 / $3.38 | 35 | 99.2% |
| 11 | DigitalOcean# | 3.81% | 10k | $1.04 / $2.09 | 52 | 100.0% |
| 12 | Azure# | 4.07% | 62k | $1.91 / $3.83 | 81 | 100.0% |
| 13 | Baidu Qianfanfp8# | 6.20% | 76k | $0.82 / $1.64 | 70 | 98.6% |
{
"provider": {
"order": [
"cloudflare",
"siliconflow/fp8",
"crusoe/fp8"
],
"ignore": [
"alibaba/fp8",
"atlas-cloud/fp4",
"digitalocean",
"azure/us",
"baidu/fp8"
]
}
}Point your agent at llms.txt
And which provider to use.
Releasing a model? Test it before launch.
of DeepSeek V4 Pro 0423's tool calls fail on Baidu Qianfan (fp8).
as often on Baidu Qianfan (fp8) as on Cloudflare, its best provider.
broken tool calls a day across its 13 providers.