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)

# pip install -U transformers accelerate
# 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, device_map="auto")
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=256)
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
```

ShuhanXue commited on 3 days ago

Commit

27317ce

verified ·

1 Parent(s): d95e30c

Add SLE v1 published evaluation results

Browse files

Add five published SLE v1 metrics (0–100) for MiniMax-M3. This submission is for M3, not M3.1-Flash.

Source: https://sciences-last-exam.org/leaderboard
Benchmark: https://huggingface.co/datasets/Gen-Verse/Science-Last-Exam
Definitions are pinned to dataset revision 022231bb779ed2b827c561b38b0902d0ea28c667. The composite equally weights four capability metrics over the 90-task scope; End-to-End Science and Hypothesis Discovery are excluded. Values retain the 1 November 2026 published chart-label precision.

Harness: Claude Code; reported effort: high. API endpoint measurements, not local or quantized weight evaluation. No verified badge is claimed. The benchmark is in private preview and native Benchmark activation is pending.

Files changed (1) hide show

.eval_results/sle-v1-20261101.yaml +50 -0

.eval_results/sle-v1-20261101.yaml ADDED Viewed

	@@ -0,0 +1,50 @@

+- dataset:
+    id: Gen-Verse/Science-Last-Exam
+    task_id: composite-score
+    revision: 022231bb779ed2b827c561b38b0902d0ea28c667
+  value: 33.0
+  source:
+    url: https://sciences-last-exam.org/leaderboard
+    name: SLE published leaderboard (1 Nov 2026)
+  notes: Claude Code; high reasoning; API endpoint evaluation; published snapshot
+    2026-11-01. Not a local-weight run.
+- dataset:
+    id: Gen-Verse/Science-Last-Exam
+    task_id: protocol-planning-mean-score
+    revision: 022231bb779ed2b827c561b38b0902d0ea28c667
+  value: 43.0
+  source:
+    url: https://sciences-last-exam.org/leaderboard
+    name: SLE published leaderboard (1 Nov 2026)
+  notes: Claude Code; high reasoning; API endpoint evaluation; published snapshot
+    2026-11-01. Not a local-weight run.
+- dataset:
+    id: Gen-Verse/Science-Last-Exam
+    task_id: experiment-observation-pass-rate
+    revision: 022231bb779ed2b827c561b38b0902d0ea28c667
+  value: 38.0
+  source:
+    url: https://sciences-last-exam.org/leaderboard
+    name: SLE published leaderboard (1 Nov 2026)
+  notes: Claude Code; high reasoning; API endpoint evaluation; published snapshot
+    2026-11-01. Not a local-weight run.
+- dataset:
+    id: Gen-Verse/Science-Last-Exam
+    task_id: scientific-computing-full-credit-rate
+    revision: 022231bb779ed2b827c561b38b0902d0ea28c667
+  value: 22.0
+  source:
+    url: https://sciences-last-exam.org/leaderboard
+    name: SLE published leaderboard (1 Nov 2026)
+  notes: Claude Code; high reasoning; API endpoint evaluation; published snapshot
+    2026-11-01. Not a local-weight run.
+- dataset:
+    id: Gen-Verse/Science-Last-Exam
+    task_id: scientific-software-operation-full-credit-rate
+    revision: 022231bb779ed2b827c561b38b0902d0ea28c667
+  value: 29.0
+  source:
+    url: https://sciences-last-exam.org/leaderboard
+    name: SLE published leaderboard (1 Nov 2026)
+  notes: Claude Code; high reasoning; API endpoint evaluation; published snapshot
+    2026-11-01. Not a local-weight run.