frames·cloud

Qwen3.6-27B

qwen-qwen3-6-27b

fields

nameREQ Qwen3.6-27B www.marktechpost.com 2026-05-01
vendorREQ Alibaba Qwen www.marktechpost.com 2026-05-01
released_at 2026-04-22 www.marktechpost.com 2026-05-01
modality text, code, image, video2 revisions www.marktechpost.com 2026-05-01
access open-weights www.marktechpost.com 2026-05-01
context_window 262144 www.marktechpost.com 2026-05-01
parameters 27B www.marktechpost.com 2026-05-01
benchmark_score 77.2 www.marktechpost.com 2026-05-01
benchmark_name SWE-Bench Verified www.marktechpost.com 2026-05-01
announcement_url https://qwen.ai/blog?id=qwen3.6-27b2 revisions qwen.ai 2026-05-01

history · 10 fields · 12 revisions

name1 revision
Qwen3.6-27B current www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
vendor1 revision
Alibaba Qwen current www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
released_at1 revision
2026-04-22 current www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
modality2 revisions
text, code, image, video current www.marktechpost.com · 2026-05-01
Qwen3.6-27B scores 77.2 on SWE-bench Verified, 59.3 on Terminal-Bench 2.0, 83.9 on LiveCodeBench v6. Native context 262,144 tokens (1M with YaRN). Natively multimodal: text, image, video.
text, code superseded www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
access1 revision
open-weights current www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
context_window1 revision
262144 current www.marktechpost.com · 2026-05-01
Qwen3.6-27B scores 77.2 on SWE-bench Verified, 59.3 on Terminal-Bench 2.0, 83.9 on LiveCodeBench v6. Native context 262,144 tokens (1M with YaRN). Natively multimodal: text, image, video.
parameters1 revision
27B current www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.
benchmark_score1 revision
77.2 current www.marktechpost.com · 2026-05-01
Qwen3.6-27B scores 77.2 on SWE-bench Verified, 59.3 on Terminal-Bench 2.0, 83.9 on LiveCodeBench v6. Native context 262,144 tokens (1M with YaRN). Natively multimodal: text, image, video.
benchmark_name1 revision
SWE-Bench Verified current www.marktechpost.com · 2026-05-01
Qwen3.6-27B scores 77.2 on SWE-bench Verified, 59.3 on Terminal-Bench 2.0, 83.9 on LiveCodeBench v6. Native context 262,144 tokens (1M with YaRN). Natively multimodal: text, image, video.
announcement_url2 revisions
https://qwen.ai/blog?id=qwen3.6-27b current qwen.ai · 2026-05-01
Official Qwen blog post for Qwen3.6-27B at qwen.ai
https://www.marktechpost.com/2026/04/22/alibaba-qwen-team-releases-qwen3-6-27b-a-dense-open-weight-model-outperforming-397b-moe-on-agentic-coding-benchmarks/ superseded www.marktechpost.com · 2026-05-01
Qwen3.6-27B: a dense open-weight model outperforming 397B MoE on agentic coding benchmarks. Released April 22, 2026.