of DeepSeek V4 Flash 0423's tool calls fail on Azure.
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 | Crusoefp8# | 0.46% | 159k | $0.14 / $0.28 | 49 | 100.0% |
| 2 | Wafermxfp4# | 0.51% | 161k | $0.60 / $0.17 | 76 | 100.0% |
| 3 | Relacefp4# | 0.58% | 848k | $0.04 / $0.45 | 85 | 100.0% |
| 4 | Fireworks# | 0.70% | 11k | $0.10 / $0.21 | 99 | 73.8% |
| 5 | DigitalOcean# | 0.73% | 64k | $0.10 / $0.20 | 33 | 100.0% |
| 6 | Cloudflare# | 0.97% | 117k | $0.44 / $1.32 | 72 | 99.6% |
| 7 | SiliconFlowfp8# | 1.06% | 131k | $0.13 / $0.28 | 51 | 99.4% |
| 8 | DeepInfrafp8# | 1.50% | 443k | $0.09 / $0.18 | 24 | 99.5% |
| 9 | GMICloudfp8# | 1.98% | 684k | $0.09 / $0.18 | 42 | 100.0% |
| 10 | Parasailfp8# | 2.32% | 140k | $0.14 / $0.28 | 69 | 100.0% |
| 11 | Alibaba Cloud Int.fp8# | 2.70% | 43k | $0.15 / $0.30 | 86 | 95.0% |
| 12 | StreamLakefp8# | 2.89% | 328k | $0.07 / $0.14 | 51 | 100.0% |
| 13 | Baidu Qianfanfp8# | 4.06% | 1.0M | $0.05 / $0.09 | 68 | 96.9% |
| 14 | AtlasCloudfp4# | 4.27% | 5.2k | $0.14 / $0.28 | 37 | 99.6% |
| 15 | Venice# | 4.33% | 38k | $0.10 / $0.19 | 24 | 100.0% |
| 16 | Azure# | 5.81% | 34k | $0.21 / $0.56 | 88 | 96.0% |
Recommended OpenRouter settings for DeepSeek V4 Flash 0423
{
"provider": {
"order": [
"crusoe/fp8",
"wafer/mxfp4",
"relace/fp4"
],
"ignore": [
"baidu/fp8",
"atlas-cloud/fp4",
"venice",
"azure/us"
]
}
}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 Azure as on Crusoe (fp8), its best provider.
broken tool calls a day across its 16 providers.