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

Libraries

How to use MiniMaxAI/MiniMax-M3 with Transformers:

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

pipe = pipeline("image-text-to-text", model="MiniMaxAI/MiniMax-M3", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("MiniMaxAI/MiniMax-M3", trust_remote_code=True)
model = AutoModelForMultimodalLM.from_pretrained("MiniMaxAI/MiniMax-M3", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MiniMaxAI/MiniMax-M3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MiniMaxAI/MiniMax-M3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MiniMaxAI/MiniMax-M3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/MiniMaxAI/MiniMax-M3

SGLang

How to use MiniMaxAI/MiniMax-M3 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 "MiniMaxAI/MiniMax-M3" \
    --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": "MiniMaxAI/MiniMax-M3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "MiniMaxAI/MiniMax-M3" \
        --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": "MiniMaxAI/MiniMax-M3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use MiniMaxAI/MiniMax-M3 with Docker Model Runner:
```
docker model run hf.co/MiniMaxAI/MiniMax-M3
```

bingran-you commited on 9 days ago

Commit

5f946db

verified ·

1 Parent(s): fa68495

Draft: Add SkillsBench v1.1 evaluation result

Browse files

Draft for @bingran-you review; maintainers please do not merge yet.

## Summary

Adds MiniMax M3's SkillsBench v1.1 with-skills evaluation result to Hugging Face eval results.

This is a metadata-only PR adding `.eval_results/skillsbench.yaml`. It does not change model files, code, dependencies, or runtime behavior.

## Result

- Benchmark: SkillsBench v1.1
- Benchmark dataset: https://huggingface.co/datasets/benchflow/skillsbench
- Task id: `skillsbench_v1_1`
- Value: `53.0`
- Mode: with-skills
- Harness: BenchFlow
- Agent: OpenHands
- Coverage: 87 tasks x 3 trials, full 261/261 selected trials
- Recomputed date: 2026-06-11

## Source

Official export:
https://huggingface.co/datasets/benchflow/skillsbench-leaderboard/raw/main/leaderboard/skillsbench/v1.1/official.json

Trajectory/result archive:
https://huggingface.co/datasets/benchflow/skillsbench-leaderboard

## Notes

SkillsBench has paired with-skills and without-skills scores. This PR submits only the with-skills score because the current Hugging Face benchmark dataset defines `skillsbench_v1_1` as the default leaderboard task. A no-skills leaderboard can be added separately if/when the benchmark dataset adds a second task id.

Files changed (1) hide show

.eval_results/skillsbench.yaml +11 -0

.eval_results/skillsbench.yaml ADDED Viewed

	@@ -0,0 +1,11 @@

+- dataset:
+    id: benchflow/skillsbench
+    task_id: skillsbench_v1_1
+    revision: be2a6ce2cb1f4ff67ce937307cade0c5a0477a13
+  value: 53.0
+  date: "2026-06-11"
+  source:
+    url: https://huggingface.co/datasets/benchflow/skillsbench-leaderboard/raw/main/leaderboard/skillsbench/v1.1/official.json
+    name: SkillsBench v1.1 official leaderboard
+    user: benchflow
+  notes: "with-skills; BenchFlow harness; OpenHands agent; 87 tasks x 3 trials; full 261/261 coverage"