of Qwen3.5-9B's tool calls fail on SiliconFlow (fp8).
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 | Darkbloomfp4# | 0.08% | 35k | $0.08 / $0.13 | 43 | 99.9% |
| 2 | Together# | 0.25% | 149k | $0.17 / $0.25 | 81 | 98.7% |
| 3 | Prime Intellectbf16# | 0.58% | 86k | $0.17 / $0.25 | 108 | 99.6% |
| 4 | DeepInfrabf16# | 0.84% | 131k | $0.10 / $0.15 | 31 | 94.2% |
| 5 | SiliconFlowfp8# | 89.86% | 162k | $0.10 / $0.15 | 23 | 96.6% |
Recommended OpenRouter settings for Qwen3.5-9B
{
"provider": {
"order": [
"darkbloom/fp4",
"together",
"primeintellect/bf16"
],
"ignore": [
"siliconflow/fp8"
]
}
}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 SiliconFlow (fp8) as on Darkbloom (fp4), its best provider.
broken tool calls a day across its 5 providers.