howtospark
models / MiniMaxAI/MiniMax-Music3
?

2.4B params
Sparks
~18 GB
~10 GB
~9 GB
~9 GB
110 usable
Spec decode
On the Sparks — 1× · 1-bit GGUF (dynamic)
spark
101 GB free
110 usable
9* / 110 GB
8% of usable
Dense weightsActivations + graphsOS reserve

* segment sizes marked with an asterisk are estimates pending a measured run

Estimated per-node footprint, weights + KV sharded across the Spark (tensor parallel). Overhead ~8 GB/node; KV at FP16.

Optimize
Verdict

2B params, ~10 GB at native precision — fits on one Spark as-is.

Make it fast

Once it fits: fewer bytes per token and fewer decode steps against 273 GB/s.

Add a speculative decoder

Decode is bandwidth-bound at ~28.1 tok/s ceiling; a trained EAGLE-style draft multiplies tokens per weight-read.

Make it yours

Change what the model does — adapt, edit, or steer it — independent of size.

Fine-tune it

LoRA/QLoRA adapts behavior on one Spark without touching the base weights.

MiniMaxAI/MiniMax-Music3 — How to Spark · How To Spark