Early previewOpenAI-compatible

GPUs go brrrr

Your models. One connection.

Try it now View API
$ gm connect --model your-next-ideaOne endpoint. Your choice of intelligence.
Model routing Familiar API Wallet connectivity Built for builders
Models

Choose a model.

Planned model families.

◎Planned

OpenAI

General

Try demo
✳Planned

Anthropic

Reasoning

Try demo
✦Planned

Google

General

Try demo
≈Planned

DeepSeek

Reasoning

Try demo
⌘Planned

Qwen

Code

Try demo
∞Planned

Meta

General

Try demo
PlaygroundINTERACTIVE DEMO

Try a prompt.

gmgpu / playgroundLOCAL PREVIEW
Input

Local demo · No credits required.

OutputREADY

Your response appears here.

YOUR APPGM ROUTERMODEL
For developers

One API.
You know the rest.

Try the request flow 
import OpenAI from "openai"; const gm = new OpenAI({  baseURL: process.env.GMGPU_BASE_URL,  apiKey: process.env.GMGPU_API_KEY,}); const response = await gm.chat.completions.create({  model: "gmgpu/fast",  messages: [{ role: "user", content: "gm, world" }],});
Preview contract · Set your gateway URL and user API key at launch.
◉

Your next idea starts here.

gm, builder.

Try it now