Skip to content
Computer vision

Qwen3.6-35B-A3B abliterated

MoE-LLM для агентного кодинга

1 min readupdated

Abliterated-вариант Qwen3.6-35B-A3B от jsantillana. Это causal language model с vision encoder: 35 млрд параметров всего и 3 млрд активных, MoE на 256 экспертов. Карточка повторяет описание базовой Qwen3.6-35B-A3B, без отдельных деталей abliteration.

Ключевые особенности

  • MoE: 256 экспертов, 8 routed + 1 shared, активируется 3 млрд параметров.
  • Контекст 262 144 токена нативно, расширяется до 1 010 000.
  • Vision encoder и задача image-text-to-text.
  • Thinking Preservation: можно сохранять reasoning из прошлых сообщений.
  • Заточен под agentic coding: frontend-воркфлоу и reasoning на уровне репозитория.

Пример

Сценарий из карточки: попросить модель разобрать frontend-воркфлоу или задачу на уровне репозитория — именно agentic coding и repository-level reasoning Qwen3.6 называет ключевыми.

ПодробнееSWE-bench Verified: 73.4 у Qwen3.6-35BA3B против 70.0 у Qwen3.5-35BA3B.
Terminal-Bench 2.0: 51.5 против 40.5.
QwenWebBench: 1397 против 978.
Данных о VRAM в карточке нет; модель совместима с Transformers, vLLM, SGLang, KTransformers.
Discuss Telegram 0 22

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 67 GB · ready files74 GB
  • FP8 / INT8recommendedweights 33 GB37 GB
  • 4-bitweights 19 GB21 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.

    # jsantillana/Qwen3.6-35B-A3B-abliterated
    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 jsantillana/Qwen3.6-35B-A3B-abliterated --local-dir ~/models/Qwen3.6-35B-A3B-abliterated
    # 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-abliterated (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

jsantillana/Qwen3.6-35B-A3B-abliterated
Author
jsantillana
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

220

Ask AI about this model

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

Comments