Alibaba releases Qwen3.6-27B dense model for coding tasks

🔍 Let's dive in

Alibaba's Qwen team released Qwen3.6-27B, a dense 27-billion-parameter open-weight model designed for coding tasks. The model reportedly surpasses the previous Qwen3.5-397B-A17B across coding benchmarks while being significantly smaller at 55.6GB versus 807GB. A blogger tested a quantized 16.8GB version locally and demonstrated code generation capabilities including SVG creation tasks.

Lead coverage: Simon Willison — Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model ↗

🕰 The timeline · 1 source

Simon Willison analyst · 1d ago · 2/5

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

Alibaba's Qwen team released Qwen3.6-27B, a dense 27-billion-parameter open-weight model designed for coding tasks. The model reportedly surpasses the previous Qwen3.5-397B-A17B across coding benchmarks while being significantly smaller at 55.6GB versus 807GB. A blogger tested a quantized 16.8GB version locally and demonstrated code generation capabilities including SVG creation tasks.

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

— Simon Willison

🏷 Tags

Llama

🔧 Debug

Cluster ID: 442f85e169
Importance (max): 2
Members: 1
Sources: Simon Willison
Earliest: 2026-04-22T16:45:23.000Z
Latest: 2026-04-22T16:45:23.000Z
Lead URL: https://simonwillison.net/2026/Apr/22/qwen36-27b