MiniMax M3 VRAM requirements

How much VRAM does MiniMax M3 need?

MiniMax M3 needs 266 GB at Q4_K_M, 466 GB at Q8_0 and 876 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.

Fine-tuning MiniMax M3? MiniMax M3 VRAM for LoRA, QLoRA and full training.

MiniMax M3 GGUF files on Hugging Face

The 8 GGUF files llama.cpp picks for MiniMax M3 with -hf <repo>:<QUANT>, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.

GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 264.0 GB, in 10⁹ bytes.

Which MiniMax M3 GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac

From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (246 GB). One request, FP16 KV cache.

Run MiniMax M3 with llama.cpp

llama-server -hf unsloth/MiniMax-M3-GGUF:Q4_K_M -c 892928

UD-Q4_K_M/MiniMax-M3-UD-Q4_K_M-00001-of-00007.gguf, 264.0 GB, from unsloth/MiniMax-M3-GGUF (checked 2026-09-29). -c 892928 is the longest Q4_K_M context on the M3 Ultra Mac Studio (512 GB, 384 GB usable). The file is 5.28 GB over the Q4_K_M estimate, so -c counts the file.

MiniMax M3 at 8K, 32K, 128K and 1M (full) tokens of context

All 60 layers use full attention. Each extra token of context adds 120 KB of FP16 cache per request. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.

How fast MiniMax M3 writes

Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth and the parameters read per token. A dash means it does not fit on one card. Try other settings in the speed calculator.

Longest context on a multi-GPU group

No single card in this table holds it at Q4_K_M, Q8_0 or FP8, so these rows use GPU groups: how many tokens of context MiniMax M3 fits on each, in one tensor-parallel group with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.

Model details

Parameters

427.0B (427,040,140,160)

Experts

128 routed experts, all loaded

Active per token

23B

Layers

All 60 layers use full attention

Attention cache

4 KV heads × 128

Context length

1,048,576 tokens

Published weights

796 GB (BF16)

On Hugging Face

MiniMaxAI/MiniMax-M3

Reported benchmarks

None published in the model card for GPQA, MMLU-Pro or LiveCodeBench.

MiniMax M3 next to similar-sized models

The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.