Skip to content
Computer vision

Qwen3.6-35B-A3B

открытая MoE-модель для агентного кодинга

1 min readupdated

Qwen3.6-35B-A3B — открытая мультимодальная LLM от Qwen с Vision Encoder. Всего 35B параметров, но в работе активируется только 3B благодаря MoE: 256 экспертов, 8 routed + 1 shared. Основной акцент — стабильность и агентный кодинг: фронтенд-воркфлоу и reasoning на уровне репозитория.

Ключевые особенности

  • 35B total / 3B activated; MoE с 256 экспертами, 8 routed + 1 shared.
  • Agentic Coding: увереннее работает с frontend-задачами и reasoning на уровне репозитория.
  • Thinking Preservation: можно сохранять reasoning из предыдущих сообщений для итеративной разработки.
  • Нативный контекст 262 144 токена, расширяется до 1 010 000.
  • Vision Encoder и task image-text-to-text; совместимость с Transformers, vLLM, SGLang, KTransformers.

Пример

Сценарий из карточки: попросить модель разобрать frontend-задачу и внести изменения на уровне репозитория, сохранив reasoning из прошлых сообщений через Thinking Preservation. Например: «Разбери репозиторий, найди причину бага во фронтенде и предложи патч».

ПодробнееБенчмарки: SWE-bench Verified — 73.4; SWE-bench Multilingual — 67.2; Terminal-Bench 2.0 — 51.5.
Архитектура: 40 слоёв, Gated DeltaNet + Gated Attention, обучение с MTP.
Требования к VRAM в карточке не указаны; заявлена совместимость с Transformers, vLLM, SGLang и KTransformers.
Discuss Telegram 0 30

Will it run?

Runs in 4-bit

Full precision does not fit, a quantised version does: quality barely changes down to 8 bits, more noticeably at 4 bits.

  • BF16weights 52 GB · ready files58 GB
  • FP8 / INT8recommendedweights 26 GB29 GB
  • 4-bitweights 15 GB17 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # inference-optimization/Qwen3.6-35B-A3B-RAZOR25
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download inference-optimization/Qwen3.6-35B-A3B-RAZOR25 --local-dir ~/models/Qwen3.6-35B-A3B-RAZOR25
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of Qwen3.6-35B-A3B-RAZOR25 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

inference-optimization/Qwen3.6-35B-A3B-RAZOR25
Author
inference-optimization
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

300

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments

Other

ANT-4B

ANT-4B