Instructions to use OsaurusAI/MiniMax-M3-Coder-Small with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OsaurusAI/MiniMax-M3-Coder-Small with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("OsaurusAI/MiniMax-M3-Coder-Small")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use OsaurusAI/MiniMax-M3-Coder-Small with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/MiniMax-M3-Coder-Small"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "OsaurusAI/MiniMax-M3-Coder-Small"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use OsaurusAI/MiniMax-M3-Coder-Small with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/MiniMax-M3-Coder-Small"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default OsaurusAI/MiniMax-M3-Coder-Small

Run Hermes

hermes

MLX LM

How to use OsaurusAI/MiniMax-M3-Coder-Small with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "OsaurusAI/MiniMax-M3-Coder-Small"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "OsaurusAI/MiniMax-M3-Coder-Small"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "OsaurusAI/MiniMax-M3-Coder-Small",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

MiniMax-M3-Coder-Small / README.md

Osaurus-AI

Add HumanEval pass@1=100% benchmark

33cc8fc verified 3 days ago

preview code

Raw

History Blame Contribute Delete

2.29 kB

	---
	license: other
	base_model: MiniMaxAI/MiniMax-M3
	tags: [mlx, jang, reap, awq, moe, code, multimodal, minimax-m3, osaurus, apple-silicon]
	pipeline_tag: text-generation
	---

	<p align="center"><img src="./osaurus-banner.png" alt="Osaurus" width="680"></p>
	<h1 align="center">MiniMax-M3-Coder-Small</h1>
	<p align="center"><b>🦖 Osaurus Exclusive — a compact JANG-quantized MiniMax-M3 coder (coding · agentic · multimodal) for Apple Silicon.</b></p>

	> ⚠️ JANG-format model — runs on Osaurus.
	> This uses the JANG quantization format (mixed-precision affine + AWQ + REAP expert pruning) and loads through Osaurus's native Swift runtime. It will NOT load with `transformers`, `vLLM`, or generic MLX loaders.

	## What is a JANG model?
	JANG is a mixed-precision quantization + packing format — per-projection affine bit widths + AWQ activation-aware scaling + REAP expert pruning — described by a `jang_config.json`. Weights stay quantized in GPU memory. Osaurus loads it through its native Swift JANG runtime on Apple Silicon.

	## Highlights
	- Smallest M3 coder — ~84 GB (the compact Osaurus build).
	- REAP45: keep 70/128 routed experts (45% pruned).
	- All-2-bit routed experts + AWQ (gate/up 2-bit AWQ-scaled, down 2-bit); attention 8-bit, shared experts 6-bit, embeddings 6-bit, lm_head 8-bit, Lightning Indexer FP16.
	- Multimodal (vision) kept.
	- Calibration: Vera (agentic-coder) + GSM8K; "floor" recipe keeps the most-salient coding experts.

	## Benchmarks
	- HumanEval: pass@1 = 100% (82/82, scrambled-half adaptive eval, seed 42; 0 failures, 0 escalations).
	- Despite 45% expert pruning + all-2-bit routed experts, coding accuracy holds at 100% — the REAP45 keep-set is a subset of the larger M3-Coder builds' proven coding experts, so coding capability is preserved while the model shrinks to ~84 GB.

	## Run it
	Load it in Osaurus (Apple Silicon) — it runs on Osaurus's native Swift JANG runtime.

	## Attribution
	- Base model: MiniMaxAI/MiniMax-M3 · Pruning: REAP (Cerebras, arXiv:2510.13999)
	- Vera calibration + testing: [@hornsman1](https://huggingface.co/hornsman1) (hornsan1 on GitHub) · math calibration: GSM8K
	- Quantization: JANG · Runtime & distribution: Osaurus