MiniMax

MiniMax M3.1 Flash

MiniMax model discovered on huggingface

Model Summary

Family

MiniMax

Version

3.1

Parameters

220.5B (est.)

Parameter counts for closed models are estimates; vendors rarely publish exact sizes.

VRAM Requirements by Quantization

Memory needed to serve MiniMax M3.1 Flash for inference, including a 20% overhead for activations and KV cache.

PrecisionVRAM neededSmallest single GPU that fits
INT4 (4-bit)123.21 GBAMD Instinct MI250X (128 GB)
INT8 (8-bit)246.43 GBAMD Instinct MI325X (256 GB)
FP16 (16-bit)492.86 GBMulti-GPU required
FP32 (32-bit)985.71 GBMulti-GPU required

Recommended GPU Configurations

Cheapest on-demand configurations to serve MiniMax M3.1 Flash at 8-bit (246 GB VRAM).

2x AMD Instinct MI250X

256 GB total VRAM · CDNA 2

~$5.00/h

16x NVIDIA T4

256 GB total VRAM · Turing · multi-node

~$8.00/h

1x AMD Instinct MI325X

256 GB total VRAM · CDNA 3

~$8.00/h

Quick GPU Planning

Use the calculator pre-filled with this exact version to estimate memory, speed, and compute requirements in a few clicks.

Access Pre-filled Calculator

Frequently Asked Questions

How much VRAM do you need to run MiniMax M3.1 Flash?

With an estimated 220.5B parameters, MiniMax M3.1 Flash needs roughly 246 GB of VRAM in 8-bit (INT8), 123 GB in 4-bit, and 493 GB in FP16, including a 20% overhead for activations and KV cache.

Which GPUs can run MiniMax M3.1 Flash?

At 8-bit quantization, the most cost-effective option is 2x AMD Instinct MI250X (256 GB combined VRAM, around $5.00/hour on-demand). Higher-end cards like the NVIDIA B200 or AMD MI355X reduce the GPU count needed.

Can MiniMax M3.1 Flash run on a single GPU?

Yes. In 8-bit, a single AMD Instinct MI325X (256 GB) fits the model.

How much does it cost to serve MiniMax M3.1 Flash in the cloud?

Renting 2x Instinct MI250X costs on the order of $5.00/hour, i.e. about $3,650/month running 24/7. Actual prices vary by provider and commitment; spot and reserved capacity can be significantly cheaper.