MiniMax-M2.7-RotorQuant-MLX-3bit

MLX 3-bit quantized variant of MiniMaxAI/MiniMax-M2.7 with RotorQuant KV-cache compression, optimized for Apple Silicon.

Overview

MiniMax-M2.7 is a massive 256-expert Mixture-of-Experts (MoE) model with 8 experts active per token, totaling approximately 456 billion parameters. This variant combines 3-bit MLX weight quantization with RotorQuant KV-cache quantization for deployment on Apple Silicon hardware.

RotorQuant applies a learned Hadamard rotation matrix to keys and values before quantization, smoothing the activation distribution for better quality retention. At 3-bit, RotorQuant's rotation-based approach is particularly valuable for preserving output quality where naive quantization would noticeably degrade.

Property	Value
Architecture	MoE (256 experts, 8 active/token)
Total Parameters	~456B
Layers	62
Hidden Size	3072
Attention Heads	48
Weight Quantization	3-bit (MLX)
KV-Cache Quantization	RotorQuant
Estimated Size	~170 GB
Base Model	MiniMaxAI/MiniMax-M2.7

Quickstart

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("majentik/MiniMax-M2.7-RotorQuant-MLX-3bit")

prompt = "What is a Comprehensive Geriatric Assessment?"
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)

response = generate(
    model,
    tokenizer,
    prompt=text,
    max_tokens=512,
)
print(response)

RotorQuant vs TurboQuant

Feature	RotorQuant	TurboQuant
Technique	Rotation-based KV quantization (Hadamard transform)	Asymmetric per-channel KV quantization
Throughput	Slightly lower throughput (rotation overhead)	Higher throughput, lower latency
Quality	Better quality preservation at low bit-widths	Good quality preservation
Best For	Quality-sensitive tasks, research	High-throughput serving, long contexts

At 3-bit quantization, RotorQuant provides meaningfully better quality than TurboQuant due to its rotation-based outlier smoothing.

Memory Estimates (Apple Silicon)

Variant	Estimated Size	Minimum Unified Memory
MLX 8-bit	~456 GB	512 GB (Mac Studio M2/M3/M4 Ultra)
MLX 5-bit	~280 GB	384 GB
MLX 4-bit	~225 GB	256 GB
MLX 3-bit	~170 GB	192 GB
MLX 2-bit	~110 GB	128 GB

Note: 3-bit quantization requires Apple Silicon with 192 GB+ unified memory, such as a Mac Studio with M2/M3/M4 Ultra.

See Also

• MiniMaxAI/MiniMax-M2.7 -- Base model
• majentik/MiniMax-M2.7-RotorQuant -- KV-cache only (transformers)
• majentik/MiniMax-M2.7-TurboQuant-MLX-3bit -- TurboQuant MLX 3-bit
• majentik/MiniMax-M2.7-RotorQuant-MLX-4bit -- MLX 4-bit
• majentik/MiniMax-M2.7-RotorQuant-MLX-2bit -- MLX 2-bit

Quant trade-off (MLX lane)

Bits	Approx size	Use case	Recommendation
2-bit	~119 GB	Aggressive quantization	Very low-RAM Macs
3-bit	~164 GB	Lossy but small	Low-RAM Macs
4-bit	~192 GB	Balanced default	Recommended for most Macs
5-bit	~228 GB	Higher fidelity	Quality-sensitive
6-bit	~274 GB	Approaching FP16 quality	High-fidelity
8-bit	~347 GB	Near-lossless reference	Fidelity-critical work

(Current variant — 3bit — is bolded.)

Variants in this family

(Showing 12 sibling variants under majentik/minimax-m2.7-*. The current variant — RotorQuant-MLX-3bit — is bolded.)

Variant	Runtime	Approx size	Use case
RotorQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
RotorQuant-MLX-2bit	mlx-lm	~885 MB	Apple Silicon, smallest
RotorQuant-MLX-3bit	mlx-lm	~1.2 GB	Apple Silicon, small
RotorQuant-MLX-4bit	mlx-lm	~1.7 GB	Apple Silicon balanced
RotorQuant-MLX-5bit	mlx-lm	~2.1 GB	Apple Silicon, higher fidelity
RotorQuant-MLX-8bit	mlx-lm	~3.2 GB	Apple Silicon reference
TurboQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
TurboQuant-MLX-2bit	mlx-lm	~885 MB	Apple Silicon, smallest
TurboQuant-MLX-3bit	mlx-lm	~1.2 GB	Apple Silicon, small
TurboQuant-MLX-4bit	mlx-lm	~1.7 GB	Apple Silicon balanced
TurboQuant-MLX-5bit	mlx-lm	~2.1 GB	Apple Silicon, higher fidelity
TurboQuant-MLX-8bit	mlx-lm	~3.2 GB	Apple Silicon reference