of Kimi K2.7 Code's tool calls fail on CoreWeave (int4).
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 | NovitaAIint4# | 0.03% | 5.9k | $0.91 / $3.84 | 23 | 100.0% |
| 2 | StreamLake# | 0.09% | 2.1k | $0.71 / $3.00 | 39 | 100.0% |
| 3 | Inceptronint4# | 0.25% | 33k | $0.56 / $3.10 | 72 | 100.0% |
| 4 | Moonshot AI# | 0.34% | 6.4k | $0.95 / $4.00 | 34 | 100.0% |
| 5 | Veniceint4# | 0.53% | 2.8k | $0.75 / $3.50 | 43 | 99.8% |
| 6 | Cloudflare# | 0.71% | 1.3k | $0.95 / $4.00 | 37 | 100.0% |
| 7 | SiliconFlowfp8# | 1.11% | 2.2k | $0.86 / $3.80 | 47 | 100.0% |
| 8 | Nebius Token Factoryfp4# | 1.28% | 3.2k | $0.95 / $4.00 | 37 | 96.4% |
| 9 | Alibaba Cloud Int.fp8# | 1.35% | 1.2k | $0.95 / $4.00 | 51 | 96.0% |
| 10 | Moonshot AIhighspeed# | 1.86% | 1.5k | $1.90 / $8.00 | 97 | 100.0% |
| 11 | CoreWeaveint4# | 3.35% | 7.4k | $0.60 / $3.00 | 36 | 98.5% |
Recommended OpenRouter settings for Kimi K2.7 Code
{
"provider": {
"order": [
"novita/int4",
"streamlake",
"inceptron/int4"
],
"ignore": [
"coreweave/int4"
]
}
}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 CoreWeave (int4) as on NovitaAI (int4), its best provider.
broken tool calls a day across its 10 providers.