MiniMax-M2.7-JANG_4M

⚠️ Requires MLX Studio to run. Standard mlx_lm cannot load mixed-precision JANG models. MLX Studio includes the JANG loader with automatic per-layer bit detection.

Follow us: X @dealignai

<p align="center"> <img src="mlx-studio-logo.png" alt="MLX Studio" width="400"/> </p> <p align="center"> <img src="jangq-logo.png" alt="JANGQ" width="200"/> </p> <p align="center"> <a href="https://vmlx.net"><img src="vmlx-app.png" alt="vMLX — run JANG models on Apple Silicon" width="820"></a> </p> <h3 align="center">⚡ All JANG models are meant to be run in <a href="https://vmlx.net">vMLX</a></h3> <div align="center">

MiniMax-M2.7 JANG_4M

MiniMax M2.7 228B MoE — 4.06-bit mixed precision, 115 GB

The standard JANG profile — best quality-to-size ratio for Apple Silicon.

</div>

Recommended: Run in MLX Studio for best experience including thinking mode support and optimized MoE inference.

Important Settings

MiniMax M2.7 is an always-reasoning model. It thinks before answering on every prompt.

Model Details

JANG_4M Bit Allocation

JANG protects routing and attention at full precision while compressing the 256 expert MLPs — where MoE models are most tolerant of quantization. The router is kept at float16 (no quantization) for maximum routing precision.

MMLU Comparison — All JANG Profiles (200q, reasoning ON)

JANG_6M not benchmarked due to slow generation (~20 tok/s). Near-lossless 6-bit expected to match or exceed 4M/3L.

Why JANG for MiniMax

Standard MLX quantization on MiniMax produces completely broken output at ALL bit levels (~25% MMLU = random guessing). JANG's mixed-precision approach is the only working quantized MiniMax on Apple Silicon.

On M2.5, JANG_2L achieved 74% MMLU vs MLX's 25% (random). M2.7 results pending.

All Quantizations

Requirements

• Apple Silicon Mac with 128 GB unified memory
• MLX framework
• MLX Studio recommended

Tool Use / Agent Mode

MiniMax M2.7 uses interleaved thinking + tool calls — it reasons inside <think> blocks, then emits tool calls in <minimax:tool_call> format. Some clients (Opencode, etc.) may strip the <think> block and miss the tool call.

For tool-use clients, set enable_thinking=False in the chat template:

text = tokenizer.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True,
    enable_thinking=False  # skips <think> injection for tool-use
)

MiniMax tool call format:

<minimax:tool_call>
<invoke name="tool_name">
<parameter name="param1">value1</parameter>
</invoke>
</minimax:tool_call>

Usage

from jang_tools.loader import load_jang_model
from mlx_lm import generate
from mlx_lm.sample_utils import make_sampler

model, tokenizer = load_jang_model("JANGQ-AI/MiniMax-M2.7-JANG_4M")
sampler = make_sampler(temp=1.0, top_p=0.95)

prompt = tokenizer.apply_chat_template(
    [{"role": "user", "content": "What is photosynthesis?"}],
    tokenize=False, add_generation_prompt=True
)
output = generate(model, tokenizer, prompt=prompt, max_tokens=2048, sampler=sampler)
print(output)

---

Support

MLX Studio | JANGQ | X @dealignai

Quantized by Jinho Jang ([email protected]) using JANG Tools v2.4.1.

---

This model is licensed under the MiniMax M2.7 Non-Commercial License. Commercial use requires prior written authorization from MiniMax ([email protected]). See LICENSE file for full terms.