Instructions to use XiaomiMiMo/MiMo-V2.6-Flash-RL with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use XiaomiMiMo/MiMo-V2.6-Flash-RL with Transformers:

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

pipe = pipeline("text-generation", model="XiaomiMiMo/MiMo-V2.6-Flash-RL", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("XiaomiMiMo/MiMo-V2.6-Flash-RL", trust_remote_code=True, device_map="auto")

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use XiaomiMiMo/MiMo-V2.6-Flash-RL with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "XiaomiMiMo/MiMo-V2.6-Flash-RL"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XiaomiMiMo/MiMo-V2.6-Flash-RL",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/XiaomiMiMo/MiMo-V2.6-Flash-RL

SGLang

How to use XiaomiMiMo/MiMo-V2.6-Flash-RL 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 "XiaomiMiMo/MiMo-V2.6-Flash-RL" \
    --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": "XiaomiMiMo/MiMo-V2.6-Flash-RL",
		"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 "XiaomiMiMo/MiMo-V2.6-Flash-RL" \
        --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": "XiaomiMiMo/MiMo-V2.6-Flash-RL",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use XiaomiMiMo/MiMo-V2.6-Flash-RL with Docker Model Runner:
```
docker model run hf.co/XiaomiMiMo/MiMo-V2.6-Flash-RL
```

Why 2.6 models performs worse in your own coding bech than older 2.5 versions?

by sarkaritamminen - opened 14 days ago

Discussion

sarkaritamminen

14 days ago

MiMo Coding Bench

All
Open
Proprietary
4 models
#	Model	Score	Size	Context	Cost	License
1	
Xiaomi
MiMo-V2.5-Pro
Xiaomi
0.737	1.0T	1.0M	$0.43 / $0.87	
2	
Xiaomi
MiMo-V2.5
Xiaomi
0.718	311B	1.0M	$0.17 / $0.34	
3	
Xiaomi
MiMo-V2.6-Pro
New
Xiaomi
0.632	1.0T	1.0M	$0.43 / $0.87	
4	
Xiaomi
MiMo-V2.6-Flash
New
Xiaomi
0.612	309B	1.0M	$0.14 / $0.28

https://llm-stats.com/benchmarks/mimo-coding-bench

Care to elaborate?

warlock-edward

13 days ago

This model performs very poorly in practice; it is all show and no substance. When I tested Qwen3.8-next-flash, deepseek-v4-flash-0731 and this model on the same task, Xiaomi’s model failed miserably. It is a model that is all show and no substance, not worth using, and it also suffers from infinite loops.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment