Instructions to use amd/MiniMax-M2.5-MXFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use amd/MiniMax-M2.5-MXFP4 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="amd/MiniMax-M2.5-MXFP4", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("amd/MiniMax-M2.5-MXFP4", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("amd/MiniMax-M2.5-MXFP4", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use amd/MiniMax-M2.5-MXFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "amd/MiniMax-M2.5-MXFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "amd/MiniMax-M2.5-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/amd/MiniMax-M2.5-MXFP4

SGLang

How to use amd/MiniMax-M2.5-MXFP4 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "amd/MiniMax-M2.5-MXFP4" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "amd/MiniMax-M2.5-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "amd/MiniMax-M2.5-MXFP4" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "amd/MiniMax-M2.5-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use amd/MiniMax-M2.5-MXFP4 with Docker Model Runner:
```
docker model run hf.co/amd/MiniMax-M2.5-MXFP4
```

Model Overview

Model Architecture: MiniMaxM2ForCausalLM
- Input: Text
- Output: Text
Supported Hardware Microarchitecture: AMD MI300 MI350/MI355
ROCm: 7.0
PyTorch: 2.8.0
Transformers: 4.57.1
Operating System(s): Linux
Inference Engine: SGLang/vLLM
Model Optimizer: AMD-Quark (v0.11)
- Weight quantization: OCP MXFP4, Static
- Activation quantization: OCP MXFP4, Dynamic

Model Quantization

The model was quantized from MiniMaxAI/MiniMax-M2.5 which was converted to bf16 using QuixiAI/MiniMax-M2.1-bf16/minimax_to_bf16.py using AMD-Quark. The weights are quantized to MXFP4 and activations are quantized to MXFP4.

Quantization scripts:

cd Quark/examples/torch/language_modeling/llm_ptq/
export exclude_layers="lm_head *block_sparse_moe.gate* *self_attn*"
python3 quantize_quark.py --model_dir $MODEL_DIR \
                          --quant_scheme mxfp4 \
                          --num_calib_data 128 \
                          --exclude_layers $exclude_layers \
                          --skip_evaluation \
                          --multi_gpu  \
                          --trust_remote_code \
                          --model_export hf_format \
                          --output_dir $output_dir

For further details or issues, please refer to the AMD-Quark documentation or contact the respective developers.

Evaluation

The model was evaluated on gsm8k benchmarks using the vllm framework.

Accuracy

Benchmark	MiniMaxAI/MiniMax-M2.5	amd/MiniMax-M2.5-MXFP4(this model)	Recovery
gsm8k (flexible-extract)	0.9401	0.9256	98.46%

Reproduction

The GSM8K results were obtained using the lm-eval framework, based on the Docker image rocm/vllm-dev:nightly.

Evaluating model in a new terminal

export VLLM_ROCM_USE_AITER=1

export model_dir=MiniMaxAI/MiniMax-M2.5-MXFP4/
log_file=minimax25-lm_eval_gsm8k_mxfp4.txt

lm_eval --model vllm --model_args pretrained=$model_dir,enforce_eager=True,trust_remote_code=True,max_model_len=16384 \
        --gen_kwargs temperature=1.0,top_p=0.95,top_k=40 \
        --tasks gsm8k  --num_fewshot 8 2>&1 | tee $log_file