Qwen3-30B-A3B-FP8

This repository contains Qwen/Qwen3-30B-A3B-FP8 together with a Furiosa Executable Bundle (FXB) for running it on FuriosaAI RNGD with Furiosa-LLM. The same model also runs on other frameworks (such as vLLM, SGLang, and Transformers); for usage with those, see the upstream Qwen/Qwen3-30B-A3B-FP8 model card.

Overview

Qwen3-30B-A3B is an auto-regressive Mixture-of-Experts (MoE) transformer with 30.5B total parameters of which about 3.3B are activated per token. It is a hybrid reasoning model that supports both thinking and non-thinking modes within a single model — switchable per request — and offers strong reasoning over text, instruction following, multilingual coverage, and tool usage. Its intended use is the same as the upstream Qwen/Qwen3-30B-A3B-FP8, and it is released under the Apache 2.0 License.

• Architecture: Qwen3-MoE (Mixture-of-Experts)
• Input / Output: Text / Text
• Supported Inference Engine: Furiosa LLM
• Supported Hardware: FuriosaAI RNGD

Quantization

Weights are quantized to FP8 (static), following the upstream FP8 release, and activations use dynamic FP8 quantization at runtime (per-token / per-block). The KV cache stays in 16-bit precision.

Features

• Reasoning. Qwen3-30B-A3B is a hybrid reasoning model: thinking mode is on by default and can be toggled per request via enable_thinking. Launch the server with --reasoning-parser qwen3 to have Furiosa-LLM parse the chain of thought into a separate field (see Basic Usage below).
• Tool calling. The model supports tool (function) calling through the hermes tool-call parser, the parser used by the Qwen3 series.

Parallelism Strategy

On RNGD, Qwen3-30B-A3B-FP8 runs with a tensor-parallel size of 32 PEs, which maps to four RNGD cards (8 PEs per card).

Usage

To run this model with Furiosa-LLM, follow the example commands below after installing Furiosa-LLM and its prerequisites.

Launch the server

Qwen3-30B-A3B-FP8 is a hybrid model that reasons by default and can switch thinking on and off (see Advanced Usage). Serve it with the qwen3 reasoning parser so the chain of thought is returned in a separate field:

furiosa-llm serve furiosa-ai/Qwen3-30B-A3B-FP8 \
  --reasoning-parser qwen3

To also enable tool (function) calling, add the hermes tool-call parser (the parser used by the Qwen3 series):

furiosa-llm serve furiosa-ai/Qwen3-30B-A3B-FP8 \
  --reasoning-parser qwen3 \
  --enable-auto-tool-choice \
  --tool-call-parser hermes

When the server is ready, you will see:

INFO:     Started server process [27507]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)

Basic Usage

The server exposes an OpenAI-compatible API. You can send a request with curl:

curl http://localhost:8000/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
    "model": "furiosa-ai/Qwen3-30B-A3B-FP8",
    "messages": [{"role": "user", "content": "What is the capital of France?"}]
    }' \
    | python -m json.tool

In thinking mode, Qwen3-30B-A3B-FP8 returns its reasoning separately from the final answer:

• response.choices[].message.reasoning (non-streaming)
• response.choices[].delta.reasoning (streaming)

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")

response = client.chat.completions.create(
    model="furiosa-ai/Qwen3-30B-A3B-FP8",
    messages=[{"role": "user", "content": "How many r's are in 'strawberry'?"}],
)

print("Reasoning:", response.choices[0].message.reasoning)
print("Answer:", response.choices[0].message.content)

Note: The reasoning field is not part of the OpenAI API specification but is a widely followed convention (the OpenAI Agents SDK, vLLM, and others). It appears only in responses that contain reasoning content; accessing it otherwise raises an AttributeError.

Advanced Usage

Toggling thinking. Qwen3-30B-A3B-FP8 reasons by default. To turn thinking off for a single request, pass enable_thinking through chat_template_kwargs; the response then carries no reasoning content, so read only message.content:

# Disable thinking for a single request
response = client.chat.completions.create(
    model="furiosa-ai/Qwen3-30B-A3B-FP8",
    messages=[{"role": "user", "content": "What is the capital of France?"}],
    extra_body={"chat_template_kwargs": {"enable_thinking": False}},
)
print(response.choices[0].message.content)

To default every request to non-thinking, launch the server with --default-chat-template-kwargs (a request can still re-enable thinking with its own chat_template_kwargs):

furiosa-llm serve furiosa-ai/Qwen3-30B-A3B-FP8 \
  --reasoning-parser qwen3 \
  --default-chat-template-kwargs '{"enable_thinking": false}'

Tool calling. With the server launched using --enable-auto-tool-choice --tool-call-parser hermes (see Launch the server), pass tools in the request and let the model decide when to call them. See the Tool Calling guide for a complete client example and details on tool-choice options.

Learn more

• Tool Calling — parsers, tool-choice options, and more examples
• Furiosa-LLM Server (furiosa-llm serve) — full OpenAI-compatible API reference and serving options
• Qwen/Qwen3-30B-A3B-FP8 — upstream model card