We haven’t updated our website in a while. We’ve grown a lot faster than we expected, and most of our time has gone into keeping up with demand.
Relace is one of the fastest-growing open-weight inference providers: 30 trillion tokens served in the last 30 days. We’re one of the top providers on OpenRouter and usually the cheapest option for the models we serve.
OpenRouter providers by daily tokens, October 5, 2026.
We think inference is the defining infrastructure problem of the next decade. Demand for tokens is growing faster than GPUs can be built, and much of the capacity that exists today sits underutilized. When the product is a commodity, the provider with the most efficient stack wins.
Models we serve:
DeepSeek V4 Pro
DeepSeek V4 Flash
DeepSeek V4.1 Flash
Kimi K3
GLM 5.3
GLM 5.3 Flash
GLM 5.2
You can see all of them on our OpenRouter page. Alternatively, you can sign up and get an API key at app.relace.ai.
If you’re spending a lot on tokens, we’d like to talk. If you have GPUs, we’d also like to talk. And if you want to work on optimizing inference at scale, we’re hiring.
Eitan, eborgnia@relace.ai
Preston, pzhou@relace.ai