Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model

Simon Willison's Blog / 4/23/2026

📰 NewsTools & Practical UsageModels & Research

共有:

Key Points

Qwen3.6-27B is a new 27B dense open-weight model that Qwen claims delivers “flagship-level” agentic coding performance, outperforming the previous open flagship Qwen3.5-397B-A17B on major coding benchmarks.
The model size is much smaller than the prior reference model in practice: Qwen3.5-397B-A17B is about 807GB on Hugging Face, while Qwen3.6-27B is 55.6GB.
A hands-on test using an 16.8GB Unsloth GGUF quantized version (Q4_K_M) with llama-server reportedly produced an outstanding result for local generation of an SVG.
The blog post includes practical setup details (llama.cpp/llama-server installation, a specific command recipe, and notes on first-run caching), making it easier for developers to try the model locally.

Simon Willison’s Weblog

Sponsored by: Honeycomb — AI agents behave unpredictably. Get the context you need to debug what actually happened. Read the blog

22nd April 2026 - Link Blog

Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model (via) Big claims from Qwen about their latest open weight model:

Qwen3.6-27B delivers flagship-level agentic coding performance, surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B (397B total / 17B active MoE) across all major coding benchmarks.

On Hugging Face Qwen3.5-397B-A17B is 807GB, this new Qwen3.6-27B is 55.6GB.

I tried it out with the 16.8GB Unsloth Qwen3.6-27B-GGUF:Q4_K_M quantized version and llama-server using this recipe by benob on Hacker News, after first installing llama-server using brew install llama.cpp:

llama-server \
    -hf unsloth/Qwen3.6-27B-GGUF:Q4_K_M \
    --no-mmproj \
    --fit on \
    -np 1 \
    -c 65536 \
    --cache-ram 4096 -ctxcp 2 \
    --jinja \
    --temp 0.6 \
    --top-p 0.95 \
    --top-k 20 \
    --min-p 0.0 \
    --presence-penalty 0.0 \
    --repeat-penalty 1.0 \
    --reasoning on \
    --chat-template-kwargs '{"preserve_thinking": true}'

On first run that saved the ~17GB model to ~/.cache/huggingface/hub/models--unsloth--Qwen3.6-27B-GGUF.

Here's the transcript for "Generate an SVG of a pelican riding a bicycle". This is an outstanding result for a 16.8GB local model:

Bicycle has spokes, a chain and a correctly shaped frame. Handlebars are a bit detached. Pelican has wing on the handlebars, weirdly bent legs that touch the pedals and a good bill. Background details are pleasant - semi-transparent clouds, birds, grass, sun.

Performance numbers reported by llama-server:

Reading: 20 tokens, 0.4s, 54.32 tokens/s
Generation: 4,444 tokens, 2min 53s, 25.57 tokens/s

For good measure, here's Generate an SVG of a NORTH VIRGINIA OPOSSUM ON AN E-SCOOTER (run previously with GLM-5.1):

Digital illustration in a neon Tron-inspired style of a grey cat-like creature wearing cyan visor goggles riding a glowing cyan futuristic motorcycle through a dark cityscape at night, with its long tail trailing behind, silhouetted buildings with yellow-lit windows in the background, and a glowing magenta moon on the right.