Arrakis / Models / MiniMax-M2.5

MiniMax-M2.5 GPU requirements

MiniMax-M2.5 has 229B parameters. To serve it, plan for about 549 GB of GPU memory in 16-bit, 274 GB in 8-bit or 137 GB in 4-bit, including about 20% for the KV cache and activations.

It is a mixture of experts: 256 experts, 8 of them active for each token. All of them stay in GPU memory, so plan memory for all 229B parameters.

Deploy this model Check another model

What it takes to serve it

GPU memory to run it for inference, with the smallest GPU setup that holds it and the lowest price on Arrakis right now.

Precision	Memory and the smallest setup
16-bit	549 GB: 4 × H200, from $13.76 per hour for all 4
8-bit	274 GB: 2 × H200, from $6.88 per hour for all 2
4-bit	137 GB: one H200, from $3.44 per GPU-hour

What it takes to fine-tune it

QLoRA trains adapters on a 4-bit frozen model, LoRA on a 16-bit one, and full fine-tuning updates every weight. Figures are before activations, which grow with batch size and sequence length.

Method	Memory and the smallest setup
QLoRA	160 GB: one B200, from $6.33 per GPU-hour
LoRA	549 GB: 4 × H200, from $13.76 per hour for all 4
Full fine-tuning	3,659 GB: more than one 8-GPU machine

About the model

From its card and files on Hugging Face. Memory figures here follow the GPU memory guide.

Parameters: 229B
Experts: 256
Experts for each token: 8
Context window (tokens): 196,608
Published in: FP8
License: other
Released: 2026-02-12
Weights: MiniMaxAI/MiniMax-M2.5

Deploy it on Arrakis

Open Models in the console, choose vLLM, SGLang or llama.cpp, and rent a GPU it fits on. You get an OpenAI-compatible endpoint with its own API key, and root SSH to the machine. Or run it yourself on any machine with enough GPU memory:

vLLM

vllm serve MiniMaxAI/MiniMax-M2.5 --tensor-parallel-size 4

Deploy this model How deploying works

Questions

Live prices for every GPU are on the GPU prices page.

How much GPU memory does MiniMax-M2.5 need?

To serve it, about 549 GB in 16-bit, 274 GB in 8-bit, 137 GB in 4-bit, including about 20% for the KV cache and activations. To fine-tune it, about 160 GB with QLoRA, 549 GB with LoRA and 3,659 GB for full fine-tuning with Adam, before activations.

Can I run MiniMax-M2.5 on one GPU?

Yes, quantized. In 4-bit it needs about 137 GB, which fits on one H200. In 16-bit it needs 4 H200 GPUs.

What does it cost to run MiniMax-M2.5?

In 16-bit it needs 4 H200 GPUs; in 4-bit, one H200. Each setup on this page shows the lowest price per GPU-hour on Arrakis right now, and you pay by the hour from a prepaid balance.

How do I deploy MiniMax-M2.5 on Arrakis?

Open Models in the Arrakis console, paste MiniMaxAI/MiniMax-M2.5, choose vLLM, SGLang or llama.cpp, and rent a GPU it fits on. You get an OpenAI-compatible endpoint with its own API key, and root SSH to the machine.

More models

Every model, and a check for any other.