of GLM 5.2's tool calls fail on Mistral.
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 | Friendli# | 0.15% | 12k | $1.40 / $4.40 | 131 | 100.0% |
| 2 | Waferfp4# | 0.16% | 525k | $0.01 / $7.00 | 98 | 100.0% |
| 3 | Morphfp8# | 0.19% | 2.8k | $0.18 / $1.54 | 45 | 97.1% |
| 4 | StreamLakefp8# | 0.29% | 96k | $0.61 / $1.92 | 48 | 100.0% |
| 5 | Inceptronfp4# | 0.33% | 36k | $1.20 / $4.40 | 73 | 99.7% |
| 6 | DeepInfrafp4# | 0.40% | 35k | $0.56 / $1.80 | 36 | 100.0% |
| 7 | Decartfp4/fast# | 0.45% | 63k | $1.20 / $8.00 | 485 | 99.9% |
| 8 | Cloudflare# | 0.68% | 4.4k | $0.19 / $8.00 | 87 | 99.9% |
| 9 | CoreWeavefp4# | 0.73% | 439k | $0.62 / $2.42 | 156 | 99.6% |
| 10 | Basetenfp8/fast# | 0.74% | 13k | $2.10 / $6.60 | 172 | 99.6% |
| 11 | Together# | 0.76% | 60k | $1.40 / $4.40 | 130 | 99.6% |
| 12 | NovitaAIfp8# | 0.83% | 24k | $0.65 / $2.04 | 50 | 100.0% |
| 13 | Parasailfp4# | 0.98% | 9.1k | $1.40 / $4.40 | 176 | 98.9% |
| 14 | Ambientfp8# | 0.99% | 12k | $0.60 / $2.00 | 57 | 99.2% |
| 15 | Nebius Token Factoryfp4# | 1.02% | 12k | $1.40 / $4.40 | 37 | 93.6% |
| 16 | Venicefp8# | 1.07% | 1.3k | $1.40 / $4.40 | 37 | 100.0% |
| 17 | Relace# | 1.09% | 39k | $0.01 / $7.91 | 84 | 100.0% |
| 18 | DigitalOcean# | 1.13% | 5.4k | $0.70 / $2.20 | 87 | 99.9% |
| 19 | Basetenfp8# | 1.59% | 32k | $1.40 / $4.40 | 101 | 100.0% |
| 20 | SiliconFlowfp8# | 1.67% | 9.3k | $0.70 / $2.20 | 82 | 99.4% |
| 21 | Baidu Qianfanfp4/fast# | 1.72% | 5.9k | $0.62 / $2.17 | 77 | 100.0% |
| 22 | Baidu Qianfanfp8# | 1.84% | 17k | $0.62 / $1.94 | 67 | 99.9% |
| 23 | Z.aifp8# | 1.89% | 20k | $1.40 / $4.40 | 59 | 100.0% |
| 24 | Alibaba Cloud Int.fp8# | 2.18% | 4.8k | $0.97 / $3.04 | 60 | 99.4% |
| 25 | Mistralnvfp4/zdr# | 2.37% | 11k | $1.40 / $4.40 | 79 | 99.4% |
| 26 | Alibaba Cloud Int.fp8/fast# | 2.63% | 3.6k | $2.31 / $7.26 | 84 | 96.6% |
| 27 | Mistral# | 5.14% | 2.4k | $1.54 / $4.84 | 104 | 100.0% |
Also ignored, under 1,000 tool calls a day: AtlasCloud (fp8) 4.23%.
Recommended OpenRouter settings for GLM 5.2
{
"provider": {
"order": [
"friendli",
"wafer",
"morph/fp8"
],
"ignore": [
"atlas-cloud/fp8",
"mistral/eu"
]
}
}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 Mistral as on Friendli, its best provider.
broken tool calls a day across its 23 providers.