MiniMax-M2.7-JANGTQ4

<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 JANGTQ4

MiniMax M2.7 228B MoE — 4-bit routed codebook + 8-bit affine, ~113 GB

Highest-quality JANGTQ profile — near-bf16 accuracy at ~50% of bf16 disk.

</div>

⚠️ Recommended: Run in MLX Studio or Osaurus. Both bundle the JANGTQ runtime (custom Metal kernels for codebook + Hadamard matmul). Stock mlx_lm.load() will NOT load this checkpoint.

Follow development on Twitter: @jangq_ai

---

What is JANGTQ4?

JANGTQ (JANG TurboQuant) is the most-compressed, highest-quality JANG quantization format. Routed expert weights stay in a compact codebook + Hadamard-rotated form at runtime — no decompression to affine — and the matmul path uses custom Metal kernels that read packed uint32 weights, look up centroids in a 16-entry codebook (at 4-bit), and accumulate dot products against a Hadamard-rotated input (QuIP# rotate-input-once math).

JANGTQ4 trades disk size for quality — 4-bit codebooks capture the routed- expert weight distribution near-losslessly. Pick JANGTQ4 when RAM is available and you want the closest quality to bf16 on Apple Silicon. Pick JANGTQ (2-bit) for the smallest footprint at minimal quality cost.

JANGTQ2 vs JANGTQ4 vs bf16

Pick JANGTQ4 when RAM is available and you want the highest-quality 4-bit MiniMax M2.7. Pick the 2-bit JANGTQ for the smallest disk / RAM footprint.

---

Model Details

JANGTQ4 Bit Allocation

Important Settings

MiniMax M2.7 is an always-reasoning model. The chat template unconditionally opens <think>\n at each assistant turn.

Strip <think>…</think> from the response before using the final answer.

---

Usage

pip install jang-tools

from huggingface_hub import snapshot_download
from jang_tools.load_jangtq import load_jangtq_model
from mlx_lm import generate

model_path = snapshot_download("JANGQ-AI/MiniMax-M2.7-JANGTQ4")
model, tokenizer = load_jangtq_model(model_path)

messages = [{"role": "user", "content": "Explain photosynthesis in 5 sentences."}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True, tokenize=False
)
out = generate(model, tokenizer, prompt, max_tokens=600, verbose=True, temperature=1.0)

Swift — MLX Studio / Osaurus

Both clients auto-detect the JANGTQ runtime from jang_config.json and route through the MiniMaxJANGTQModel class. Just load the repo — no extra flags.

---

What's In This Repo

Parser Capabilities (Tier-1 auto-detected by vmlx)

{
  "reasoning_parser": "qwen3",
  "tool_parser": "minimax",
  "think_in_template": true,
  "supports_tools": true,
  "supports_thinking": true,
  "family": "minimax_m2",
  "modality": "text",
  "cache_type": "kv"
}

<think>...</think> and <tool_call>...</tool_call> are non-special tokens by design — the application layer parses them. vmlx's CapabilityDetector reads this block verbatim and wires the qwen3 reasoning parser + minimax tool parser automatically, so streamed responses route reasoning_content and tool_calls into the OpenAI-compatible SSE fields instead of leaking into content.

License

MiniMax non-commercial (inherits from upstream — see LICENSE).

Credits

Created by Jinho Jang — [email protected]

Based on MiniMaxAI's MiniMax M2.7. JANGTQ quantization © JANGQ-AI.

Kernels: hadamardRotate, fusedGateUpSwiGLU (P17 OPT=10), gatherTQ, compiled-router math (P15) — all in jang_tools/turboquant/.