Instructions to use Qwen/Qwen3.8-Flash-Next with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Qwen/Qwen3.8-Flash-Next with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="Qwen/Qwen3.8-Flash-Next") 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("Qwen/Qwen3.8-Flash-Next") model = AutoModelForMultimodalLM.from_pretrained("Qwen/Qwen3.8-Flash-Next", 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
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Qwen/Qwen3.8-Flash-Next with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Qwen/Qwen3.8-Flash-Next" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Qwen/Qwen3.8-Flash-Next", "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/Qwen/Qwen3.8-Flash-Next
- SGLang
How to use Qwen/Qwen3.8-Flash-Next 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 "Qwen/Qwen3.8-Flash-Next" \ --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": "Qwen/Qwen3.8-Flash-Next", "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 "Qwen/Qwen3.8-Flash-Next" \ --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": "Qwen/Qwen3.8-Flash-Next", "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 Qwen/Qwen3.8-Flash-Next with Docker Model Runner:
docker model run hf.co/Qwen/Qwen3.8-Flash-Next
Quick Thanks
#90 opened 1 day ago
by
LittleVoidGames
Add modality maturity index (MMI) results in .eval_results/
#89 opened 8 days ago
by
dieuwkehupkes
Add MDPBench evaluation results
#88 opened 8 days ago
by
Delores-Lin
Inconsistent video token expansion across Transformers versions in Qwen3.8-Flash-Next
1
#87 opened 8 days ago
by
limboy058
Retokenization drift causes partial KV-cache invalidation
#86 opened 9 days ago
by
y-korobko
Keep the pre-filled thinking block out of the generation tags
#85 opened 9 days ago
by
Rocketknight1
AMD Ryzen AI Max+ PRO 495 (192GB) – Qwen Q6, 249K Context Benchmark, ROCm and Memory Results
12
#84 opened 13 days ago
by
UgurDiskaya
NVFP4 Shootout (single DGX Spark) + Qwen3.8-27B comparison
❤️ 2
#83 opened 13 days ago
by
Rieker
Notes from an independent NVFP4 port of Qwen3.8-Flash-Next
#82 opened 14 days ago
by
alexmv2025
Add IFBench evaluation result
2
#80 opened 16 days ago
by
SaylorTwift
Qwen3.8-Flash-Next at 57.6 GB: 80.5 tok/s on 2× 7900 XT, with only 32 GB of RAM
#79 opened 23 days ago
by
SkyIsNotGreen
Measured: Qwen models on Strix Halo mini PCs — 43 of our 72 published benchmark records are Qwen
#78 opened 26 days ago
by
geekom-ai-admin
support for Flash Attention 2 attention
#77 opened 28 days ago
by
mrgrandsky
Accept consecutive leading system messages in chat template
👍 1
1
#76 opened about 1 month ago
by
jlapenna
Chat template rejects consecutive leading system messages ("System message must be at the beginning.")
#75 opened about 1 month ago
by
jlapenna
Just wanted to say thanks!
➕🔥 37
#73 opened about 1 month ago
by
prevod
Update README.md
#72 opened about 2 months ago
by
Manusagents
How to limit thinking?
4
#71 opened about 2 months ago
by
MarbuelB
Strix Halo teaser: 88 GB Qwen3.8-Flash-Next (512 experts) at 45–50 tok/s on a Radeon 8060S — what we've been cooking in our llama.cpp fork
❤️👍 3
3
#70 opened about 2 months ago
by
Fbrbiasi
Qwen3.8-Flash-Next hallucinates webfetch URLs
👀 2
3
#69 opened about 2 months ago
by
g4bw
Qwen3.8 Flash Next and 27B intermittently invent unrelated user tasks and misread immediate tool results after tool calls
👀➕ 4
2
#68 opened about 2 months ago
by
drcolekat
Leave it running in opencode for the whole day and come back to something useful!
#67 opened about 2 months ago
by
catplusplus
Request: DOI
#66 opened about 2 months ago
by
mohamedmega
Qwen 3.8 flash next, vram requirements
👍 1
4
#65 opened about 2 months ago
by
mzubairtahir000714
Engram weights for domain specialization?
1
#64 opened about 2 months ago
by
AlexAlex321
Upload без звука гот.mp4
#63 opened about 2 months ago
by
AnnaYans
Model cannot SPLIT TENSOR and will not work in consumer multi gpu
😔 1
3
#62 opened 2 months ago
by
Dr-Ojinsky
Update README.md
#61 opened 2 months ago
by
daniilkolbasenko
Qwen 3.8 Flash Next VRAM requirements
1
#60 opened 2 months ago
by
yash-711
Gated DeltaNet 2
#58 opened 2 months ago
by
ancc
Identified serious issue with Flash-Next hallucinating user instructions!
1
#57 opened 2 months ago
by
laserbeans
upload 100%
#56 opened 2 months ago
by deleted
LLM TOOLS
#55 opened 2 months ago
by
GopiKrishnaReddy
Request for Qwen4-35B-A3B-Distill-FlashNext
👍 2
3
#54 opened 2 months ago
by
userr99
FYI, Qwen3.8-Flash-Next-NVFP4 works quite well on a single DGX Spark (GB10)!
4
#53 opened 2 months ago
by
donghwan-shin
multi gpu & buffer size issues is fixed please check this pr
#52 opened 2 months ago
by
gopi87
Just the best model 💘 as always Qwen Team is number 1 🥰 .
👍🤗 6
4
#51 opened 2 months ago
by
auf1r2
Premature empty turns after tool results: "No response requested." (NVFP4/SGLang)
1
#50 opened 2 months ago
by
ProCreations
Old Macs M1-M3 > Run Qwen 3.8 Flash Next from SSD at 5 tok/s
❤️ 1
#48 opened 3 months ago
by
Pummelchen
Our SGLang config + measured benchmarks on 2x RTX PRO 6000 (SM120, NVFP4)
1
#47 opened 3 months ago
by
bureksirovic
New AI = Intelligence + Knowledge
👍 1
4
#43 opened 3 months ago
by
Duonglv
I beg you for qwen3.8 42B A5b
👍👀 5
1
#42 opened 3 months ago
by
TAOTAO777
Add Toolathlon-Verified evaluation result
#41 opened 3 months ago
by
SaylorTwift
Ttu
#40 opened 3 months ago
by
Adagannswi
Add ExtractBench evaluation results
❤️ 1
#38 opened 3 months ago
by
boyang-runllama
anyone has claude code fixed template?
2
#37 opened 3 months ago
by
gopi87
Changing reasoning effort requires rebuilding whole kv cache
👍 1
2
#36 opened 3 months ago
by
paragon-of-brah
Good but ...
👍 3
3
#35 opened 3 months ago
by
jbourny
Request 9B, 35B A3B, 4B qwen 3.8 models
🔥❤️ 31
2
#34 opened 3 months ago
by
usernamebetter