Project AURIONAI Use-Cases Research & Innovation
multimodal LLM

Qwen3.8-27B

Dense 27B open-weight Qwen3.8 model, newest small open agentic coder (Aug 2026).

Alibaba (Qwen)Alibaba (Qwen)ChinaQwen3.8Family
Alibaba (Qwen)
Levelmodel
Classmultimodal LLM
Parameters27B
Context262,144
ReleasedAug 2026
Opennessopen weights permissive
Runs onconsumer GPU
Statusactive

Benchmark Results

vendor-reported
GPQA Diamondknowledge and reasoning
vendor-reported
89.2
050100

Score in %. Higher is better.

huggingface.co
LiveCodeBench v6coding
vendor-reported
90.3
050100

Score in %. Higher is better.

huggingface.co
LiveCodeBenchcoding
LiveCodeBench v6vendor-reported
90.3
050100

Score in %. Higher is better.

huggingface.co
SWE-bench Procoding
vendor-reported
61.7
050100

Score in %. Higher is better.

huggingface.co
vendor-reported
61.7
050100

Score in %. Higher is better.

huggingface.co
Terminal-Bench 2.1agents and tool use
Terminusvendor-reported
73
050100

Score in %. Higher is better.

huggingface.co
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

R&D Topics

Agentic software engineeringproduced byAgents and tool useproduced byOpen-weight frontier releases as research substrateused by

Identity

Developer
Alibaba (Qwen)
Family
Qwen3.8
Released
Aug 2026 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
hybrid Gated DeltaNet + Gated Attention, dense, MTP, vision encoder
Total parameters
27B sources
Active parameters
–
Checkpoint parameters
–
Note
1M with YaRN
Context length
262,144 tokens
Input
text, image, video
Output
text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights permissive
Licence
Apache License 2.0 sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
consumer GPU
Footprint
T4 workstation estimate (derived from parameter count, unverified). 262K native context, 1M extended.
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
Not yet on SEA-HELM
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Lineage

The notes record no lineage for this model.

Industry Use Cases

No industry use case links this model.

Sources

  1. Qwen/Qwen3.8-27B , Hugging Face huggingface.co https://huggingface.co/Qwen/Qwen3.8-27B
  2. Qwen (Qwen) huggingface.co https://huggingface.co/Qwen
  3. Text Arena: LLM Rankings & AI Model Benchmarks arena.ai https://arena.ai/leaderboard/chat/text
  4. Open-Source LLMs 2026: Which Are Actually Open-Licensed codersera.com https://codersera.com/blog/open-source-llms-landscape-2026/
  5. huggingface.co/api/models/Qwen/Qwen3.8-27B https://huggingface.co/api/models/Qwen/Qwen3.8-27B
  6. alibabacloud.com/blog/alibaba-unveils-qwen3-8-27b-and-releases... https://www.alibabacloud.com/blog/alibaba-unveils-qwen3-8-27b-and-releases-weights-of-qwen3-8-flagship-model_603463

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Qwen3.8-27B", https://aurion.semai.id/models/alibaba-qwen3-8-27b/, records of October 2026.