Model catalogMiniMax
MiniMax M3
Very long-context reasoning, up to a 1M-token window. Served with vLLM behind your own private, OpenAI-compatible /v1 endpoint — billed by the second, terminate anytime.
428B MoE · 23B act · FP8 · 1M context · 8× 80 GB · per-second billing
01Specifications
The specs the launcher enforces.
Architecture, served precision, max context, and the minimum GPU shape — the numbers that decide how MiniMax M3 runs and what it costs.
| Spec | Value |
|---|---|
| Architecture | 428B MoE · 23B act |
| Served precision | FP8 |
| Max context | 1M |
| Minimum GPU shape | 8× 80 GB |
Runs on 8× 80 GB · billed per second — final cost depends on the GPU shape.
Full pricing →02How it runs
A private endpoint, not a shared API.
Launching MiniMax M3 runs the model on your own instance and gives you a private, OpenAI-compatible /v1 endpoint and API key. Point any OpenAI-compatible SDK at your base URL, change the model name, and existing code works. The endpoint is yours alone — terminate the instance and it goes away with it.
Billing is by the second on the 8× 80 GB shape the launcher enforces, from boot to terminate. No commitment and no tier gates.
03Get started
Launch MiniMax M3 now.
One private OpenAI-compatible endpoint, one key, billed by the second. The console opens with MiniMax M3 preselected.