of Gemma 4 26B A4B's tool calls fail on AkashML (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 | NextBitbf16# | 0.00% | 67k | $0.08 / $0.26 | 74 | 100.0% |
| 2 | Darkbloom# | 0.00% | 16k | $0.04 / $0.22 | 40 | 99.6% |
| 3 | Prime Intellectbf16# | 0.02% | 55k | $0.10 / $0.30 | 50 | 100.0% |
| 4 | NovitaAIbf16# | 1.02% | 15k | $0.13 / $0.40 | 195 | 97.9% |
| 5 | io.netbf16# | 1.06% | 3.2k | $0.09 / $0.31 | 98 | 100.0% |
| 6 | DeepInfrafp8# | 1.20% | 144k | $0.07 / $0.34 | 71 | 99.9% |
| 7 | Makora# | 2.48% | 29k | $0.07 / $0.24 | 150 | 99.9% |
| 8 | DekaLLMbf16# | 3.79% | 25k | $0.06 / $0.33 | 82 | 99.7% |
| 9 | Google Vertex# | 5.30% | 2.8k | $0.15 / $0.60 | 24 | 95.5% |
| 10 | Cloudflare# | 7.74% | 31k | $0.10 / $0.30 | 49 | 100.0% |
| 11 | AkashMLfp4# | 8.70% | 1.9k | $0.07 / $0.24 | 86 | 98.8% |
Recommended OpenRouter settings for Gemma 4 26B A4B
{
"provider": {
"order": [
"nextbit/bf16",
"darkbloom",
"primeintellect/bf16"
],
"ignore": [
"dekallm/bf16",
"google-vertex/global",
"cloudflare",
"akashml/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 AkashML (fp4) as on NextBit (bf16), its best provider.
broken tool calls a day across its 11 providers.