FlagRelease/Qwen3.6-35B-A3B-nomtp-metax-FlagOS-Express

Introduction

Qwen3.6-35B-A3B is a fully open-source sparse MoE model (35B total parameters / 3B active parameters) that excels at agentic coding, significantly outperforming its predecessor Qwen3.5-35B-A3B and holding its own against dense models such as Qwen3.5-27B and Gemma4-31B. Key features include:

• Outstanding agentic coding capabilities, comparable to much larger models
• Strong multimodal perception and reasoning abilities

Integrated Deployment

• Out-of-the-box inference scripts with pre-configured hardware and software parameters
• Released FlagOS-Metax container image supporting deployment within minutes

Consistency Validation

• Rigorously evaluated through benchmark testing: Performance and results from the FlagOS software stack are compared against native stacks on multiple public.

Evaluation Results

Benchmark Result

Metrics	Qwen3.6-35B-A3B-nomtp-Nvidia-Origin	Qwen3.6-35B-A3B-nomtp-Metax-FlagOS
GPQA_Diamond	0.8283	0.803
ERQA	0.5875	0.6

Performance Benchmark Result

Metric	1k&1k 64 Concurrency	4k&1k 64 Concurrency	16k&1k 64 Concurrency	64k&1k 64 Concurrency
Equal Computing Power Ratio (FlagOS/H100)	136.94%	125.03%	113.468%	87.97%

User Guide

Environment Setup

Item	Version
Docker Version	Docker version 27.5.1, build 27.5.1-0ubuntu3~22.04.2
Operating System	Ubuntu 22.04.5 LTS (Jammy Jellyfish)

Operation Steps

Download FlagOS Image

docker pull harbor.baai.ac.cn/flagrelease-public/qwen36-35b-a3b-metax001-gems5.0.2-tree0.5.1-cxnone-plugin0.2.0-vllm0.20.2-cp312-pt28-maca37-x64-3.8.1:202607280210

Download Open-source Model Weights

pip install modelscope
modelscope download --model FlagRelease/Qwen3.6-35B-A3B-nomtp-metax-FlagOS-Express --local_dir /data/Qwen3.6-35B-A3B-nomtp

Start the Container

docker run -itd \
    --name flagos \
    --privileged \
    --ipc=host \
    --network=host \
    --shm-size 64g \
    -v /data:/data \
    -w /workspace \
    harbor.baai.ac.cn/flagrelease-public/qwen36-35b-a3b-metax001-gems5.0.2-tree0.5.1-cxnone-plugin0.2.0-vllm0.20.2-cp312-pt28-maca37-x64-3.8.1:202607280210 \
    /bin/bash
docker exec -it flagos /bin/bash

Start the Server

Start Service and Test 1k, 4k Scenarios

# Test 1k, 4K Input Scenario
export USE_FLAGGEMS=1
export VLLM_USE_FLAGGEMS=1
export FLAGGEMS_ATEN_SHAPE_AWARE=0
export VLLM_FL_FLAGOS_WHITELIST=add,cat,cos,cumsum,embedding,full,gather,le,lt,mul,pow_scalar,resolve_conj,sin,softmax,sub,zeros,zeros_like
export VLLM020_CONTIGUOUS_SINGLE_PREFILL=0
export VLLM020_GDN_PACKED_MIN_BATCH=32
export VLLM020_MOE_EVEN_K_FASTPATH=1
export VLLM020_GDN_T1_FUSED_H_SPLIT=1

vllm serve /data/Qwen3.6-35B-A3B-nomtp \
    --host 0.0.0.0 \
    --port 8020 \
    --served-model-name qwen36-moe \
    --tensor-parallel-size 2 \
    --trust-remote-code \
    --dtype bfloat16 \
    --max-model-len 67584 \
    --max-num-batched-tokens 8192 \
    --max-num-seqs 64 \
    --gpu_memory_utilization 0.9 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY","max_cudagraph_capture_size":64}' \
    --generation-config vllm \
    --no-enable-prefix-caching \
    > /workspace/qwen36-moe-server.log 2>&1 &

Start Service and Test 16k Scenarios

# Test 16k Input Scenario
export USE_FLAGGEMS=1
export VLLM_USE_FLAGGEMS=1
export FLAGGEMS_ATEN_SHAPE_AWARE=0
export VLLM_FL_FLAGOS_WHITELIST=add,cat,cos,cumsum,embedding,full,gather,le,lt,mul,pow_scalar,resolve_conj,sin,softmax,sub,zeros,zeros_like
export VLLM020_CONTIGUOUS_SINGLE_PREFILL=1
export VLLM020_GDN_PACKED_MIN_BATCH=64
export VLLM020_MOE_EVEN_K_FASTPATH=0
export VLLM020_GDN_T1_FUSED_H_SPLIT=0

vllm serve /data/Qwen3.6-35B-A3B-nomtp \
    --host 0.0.0.0 \
    --port 8020 \
    --served-model-name qwen36-moe \
    --tensor-parallel-size 2 \
    --trust-remote-code \
    --dtype bfloat16 \
    --max-model-len 67584 \
    --max-num-batched-tokens 32768 \
    --max-num-seqs 64 \
    --gpu_memory_utilization 0.9 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY","max_cudagraph_capture_size":64}' \
    --generation-config vllm \
    --no-enable-prefix-caching \
    > /workspace/qwen36-moe-server.log 2>&1 &

Start Service and Test 64k Scenarios

# Test 64k Input Scenario
export USE_FLAGGEMS=1
export VLLM_USE_FLAGGEMS=1
export FLAGGEMS_ATEN_SHAPE_AWARE=0
export VLLM_FL_FLAGOS_WHITELIST=add,cat,cos,cumsum,embedding,full,gather,le,lt,mul,pow_scalar,resolve_conj,sin,softmax,sub,zeros,zeros_like
export VLLM020_CONTIGUOUS_SINGLE_PREFILL=1
export VLLM020_GDN_PACKED_MIN_BATCH=64
export VLLM020_MOE_EVEN_K_FASTPATH=0
export VLLM020_GDN_T1_FUSED_H_SPLIT=0

vllm serve /data/Qwen3.6-35B-A3B-nomtp \
    --host 0.0.0.0 \
    --port 8020 \
    --served-model-name qwen36-moe \
    --tensor-parallel-size 2 \
    --trust-remote-code \
    --dtype bfloat16 \
    --max-model-len 67584 \
    --max-num-batched-tokens 16384 \
    --max-num-seqs 64 \
    --gpu_memory_utilization 0.9 \
    --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY","max_cudagraph_capture_size":64}' \
    --generation-config vllm \
    --no-enable-prefix-caching \
    > /workspace/qwen36-moe-server.log 2>&1 &

Service Invocation

Invocation Script

curl http://localhost:8020/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen36-moe",
    "messages": [{"role": "user", "content": "hi"}]
  }'

AnythingLLM Integration Guide

1. Download & Install

• Visit the official site: https://anythingllm.com/
• Choose the appropriate version for your OS (Windows/macOS/Linux)
• Follow the installation wizard to complete the setup

2. Configuration

• Launch AnythingLLM
• Open settings (bottom left, fourth tab)
• Configure core LLM parameters
• Click "Save Settings" to apply changes

3. Model Interaction

• After model loading is complete:
• Click "New Conversation"
• Enter your question (e.g., “Explain the basics of quantum computing”)
• Click the send button to get a response

Technical Overview

FlagOS is a fully open-source system software stack designed to unify the "model–system–chip" layers and foster an open, collaborative ecosystem. It enables a “develop once, run anywhere” workflow across diverse AI accelerators, unlocking hardware performance, eliminating fragmentation among vendor-specific software stacks, and substantially lowering the cost of porting and maintaining AI workloads. With core technologies such as the FlagScale, together with vllm-plugin-fl, distributed training/inference framework, FlagGems universal operator library, FlagCX communication library, and FlagTree unified compiler, the FlagRelease platform leverages the FlagOS stack to automatically produce and release various combinations of <chip + open-source model>. This enables efficient and automated model migration across diverse chips, opening a new chapter for large model deployment and application.

FlagGems

FlagGems is a high-performance, generic operator libraryimplemented in Triton language. It is built on a collection of backend-neutralkernels that aims to accelerate LLM (Large-Language Models) training and inference across diverse hardware platforms.

FlagTree

FlagTree is an open source, unified compiler for multipleAI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-repository multi-backend support. Forupstream model users, it provides unified compilation capabilities across multiple backends; for downstream chip manufacturers, it offers examples of Triton ecosystem integration.

FlagScale and vllm-plugin-fl

Flagscale is a comprehensive toolkit designed to supportthe entire lifecycle of large models. It builds on the strengths of several prominent open-source projects, including Megatron-LM and vLLM, to provide a robust, end-to-end solution for managing and scaling large models. vllm-plugin-fl is a vLLM plugin built on the FlagOS unified multi-chip backend, to help flagscale support multi-chip on vllm framework.

FlagCX

FlagCX is a scalable and adaptive cross-chip communication library. It serves as a platform where developers, researchers, and AI engineers can collaborate on various projects, contribute to the development of cutting-edge AI solutions, and share their work with the global community.

FlagEval Evaluation Framework

FlagEval is a comprehensive evaluation system and open platform for large models launched in 2023. It aims to establish scientific, fair, and open benchmarks, methodologies, and tools to help researchers assess model and training algorithm performance. It features:

• Multi-dimensional Evaluation: Supports 800+ modelevaluations across NLP, CV, Audio, and Multimodal fields,covering 20+ downstream tasks including language understanding and image-text generation.
• Industry-Grade Use Cases: Has completed horizonta1 evaluations of mainstream large models, providing authoritative benchmarks for chip-model performance validation.

Contributing

We warmly welcome global developers to join us:

1. Submit Issues to report problems
2. Create Pull Requests to contribute code
3. Improve technical documentation
4. Expand hardware adaptation support

License

The model weights are derived from Qwen/Qwen3.6-35B-A3B-nomtp and are open‑sourced under the Apache License 2.0: https://www.apache.org/licenses/LICENSE-2.0.txt