New Multimodal

Qwen3.8 Flash Next

Qwen's vision-language MoE model for reasoning, coding, and tool use on Jetson Thor, with NVFP4 quantization and MTP speculative decoding

Command to Run on Jetson Model Details

Parameters 125B MoE / 6B activated

Modalities

Text Image Video

Context Length 262K

License Qwen Community 1.0

Precision

NVFP4

Serve the model

Start server

Choose module, then engine and optional parameters on the left, then copy the serve command by clicking the button on the right.

Command

Call the model over Web API

Copy a client command below and paste it into your terminal to make a Web API request to the model you just served.

curl -s http://${JETSON_HOST}:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen3.8-Flash-Next",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from openai import OpenAI

client = OpenAI(
    base_url="http://${JETSON_HOST}:8000/v1",
    api_key="not-needed",  # vLLM / llama.cpp typically do not enforce a key
)

completion = client.chat.completions.create(
    model="Qwen3.8-Flash-Next",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(completion.choices[0].message.content)

Model Details

View on HuggingFace

Qwen3.8 Flash Next is Qwen’s vision-language mixture-of-experts model for reasoning, coding, and tool use. It has 125B language-model parameters with 6B activated per token, plus 51B n-gram embedding parameters. Its native context window is 262,144 tokens.

Modalities

Input: Text, image, and video

Output: Text

What makes Flash Next different

Flash Next is an early preview of Qwen’s upcoming Qwen4 architecture. It uses sparse attention to reduce the work needed to process long context, alongside n-gram embeddings that add model capacity with relatively little extra computation.

The model focuses on multi-step tasks that involve planning, calling tools, and acting on their results. Its preserved thinking keeps earlier reasoning available across conversation turns, helping it follow through on longer tasks. Thinking is enabled by default and can be disabled for direct responses.