Voxtral-Mini-3B-2507-TurboQuant-MLX-4bit

4-bit MLX weight-quantized build of mistralai/Voxtral-Mini-3B-2507 with a TurboQuant KV-cache profile. Balanced point between quality and memory for Apple Silicon.

Overview

• Base: mistralai/Voxtral-Mini-3B-2507 — 3B speech-understanding model
• Capabilities: transcription, speech translation, audio QA
• Weight precision: 4-bit (group-wise)
• KV-cache profile: TurboQuant (per-head static calibration)
• Approx. on-disk size: ~1.5 GB
• Runtime: MLX on Apple Silicon

Quickstart

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("majentik/Voxtral-Mini-3B-2507-TurboQuant-MLX-4bit")

prompt = tokenizer.apply_chat_template(
    [{"role": "user", "content": [{"type": "audio", "path": "sample.wav"},
                                  {"type": "text", "text": "Transcribe this."}]}],
    add_generation_prompt=True,
)
print(generate(model, tokenizer, prompt=prompt, max_tokens=256))

Model specs

Field	Value
Parameters	3B
Weight bits	4
Group size	64
Cache profile	TurboQuant
Size on disk	~1.5 GB
Target hardware	Apple Silicon (M1/M2/M3/M4)
License	Apache 2.0

RotorQuant vs TurboQuant

	TurboQuant	RotorQuant
Strategy	Per-head static calibration	Rotational online re-basis
Memory reduction	~3.5x on KV-cache	~4x on KV-cache
Best for	Batch transcription	Streaming / code-switching

Recommended default for most Apple-Silicon deployments. Pick 8-bit for highest fidelity, 2-bit for lightest footprint.

See also

• majentik/Voxtral-Mini-3B-2507-TurboQuant-MLX-8bit
• majentik/Voxtral-Mini-3B-2507-TurboQuant-MLX-2bit
• majentik/Voxtral-Mini-3B-2507-RotorQuant-MLX-4bit
• majentik/Voxtral-Mini-3B-2507-TurboQuant — KV-cache-only bundle
• mistralai/Voxtral-Mini-3B-2507 — upstream base model