Skip to content
EN ES
Alibaba

Qwen3.6-27B

Released 2026-04 · hybrid · 262K tokens · 4 benchmarks · Open weight
HuggingFaceGitHub

Editorial notes

Lanzado 22 abril 2026. Modelo dense 27B que SUPERA a Qwen3.5-397B-A17B (MoE, 14.7x mas params totales) en agentic coding. SWE-bench Verified a 3.7 pts de Claude Opus 4.6 (80.8). 64 capas, hidden 5120. Apache 2.0. Context 1M con YaRN.

Spec sheet

Company
Alibaba
Country
CN
Type
hybrid
Release
2026-04
Context
262K tokens
Params total
27B
License
apache-2.0
Quants
BF16, Q8_0, Q5_K_M, Q4_K_M, Q3_K_M
Slug
qwen3-6-27b

Quick install

2 tools
ollama.com 
ollama run qwen3.6:27b

Benchmarks (4)

Coding 4

Cite this model
BibTeX · APA

BibTeX 

@misc{frontier-qwen3-6-27b,
  title  = {Qwen3.6-27B},
  author = {{Alibaba}},
  year   = {2026},
  note   = {Frontier Benchmarks AI atlas. Accessed 2026-06-17},
  url    = {https://frontierbenchmarks.com/models/qwen3-6-27b}
}

APA

Alibaba (2026). Qwen3.6-27B [Large language model]. Frontier Benchmarks AI. Retrieved 2026-06-17, from https://frontierbenchmarks.com/models/qwen3-6-27b

Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.

Battle vs another model ← All models More from Alibaba Methodology