Models / MiniMax M3
Best home rig for MiniMax M3
Legacy model428B total · 23B active1 measured resultfastest plain decode: 28.9 tok/s model info ↗
Real, measured generation speeds for MiniMax M3 on home hardware, fastest first. Different quants/engines aren't directly comparable.
| Build | Quant | Engine | tok/s | Cost | Source |
|---|---|---|---|---|---|
| Mac Studio M3 Ultra 512 GB running MiniMax M3 (4-bit MLX) | 4-bit MLX | oMLX v0.4.4rc2 | 28.9 | omlx.ai ↗ |
Memory math: how much VRAM / RAM does MiniMax M3 need?
MiniMax M3 has 428B parameters, about 23B active per token (source: Hugging Face model card: ~428B total, ~23B activated). Weight size = parameters × bits-per-weight ÷ 8, before context (KV cache) and runtime overhead.
| Format | Bits/weight | Weights | Smallest common memory size that holds the weights alone |
|---|---|---|---|
| Q8_0 | 8.50 | 454.8 GB | 512 GB |
| Q6_K | 6.56 | 351.1 GB | 512 GB |
| Q5_K_M | 5.70 | 305.1 GB | 512 GB |
| Q4_K_M | 4.89 | 261.9 GB | 512 GB |
| Q3_K_M | 4.00 | 213.8 GB | 256 GB |
| Q2_K | 3.16 | 169.0 GB | 192 GB |
Bits-per-weight for GGUF quants are llama.cpp's own figures (llama.cpp quantize README). "As measured" rows apply the source's stated bits-per-weight to the parameter count; real files differ slightly because some layers stay at higher precision. Try other combinations in the VRAM calculator.
Why it can still run on smaller GPUs: it is a mixture-of-experts model, so only about 23B of its 428B parameters are used for each generated token. Runtimes such as llama.cpp (--n-cpu-moe), ExLlamaV3 or Project Maya can keep most expert weights in system RAM and only the shared layers and hot experts in VRAM. Then you need VRAM + system RAM ≥ weights, and speed is limited mainly by RAM bandwidth.
FAQ
How much VRAM does MiniMax M3 need?
The weights are about 261.9 GB at Q4_K_M (428B params × 4.89 bits ÷ 8, before context/KV cache). Add memory for context (KV cache) and the runtime. The smallest-memory setup with a measured result here, fully in memory, is Apple M3 Ultra (512 GB) at 28.9 tok/s.
What is the fastest measured hardware for MiniMax M3?
Apple M3 Ultra: 28.9 tok/s generation (oMLX v0.4.4rc2, 4-bit MLX), from https://omlx.ai/benchmarks/performance/bebzj6ar.
Related: what GPU do you need for 2026 models · all builds