Qwen 3.8 27B

Open Source

Open Weights

Parameters

27B

Context Length

262K

Modality

Text

Architecture

Dense

License

Apache 2.0

Release Date

5 Aug 2026

Knowledge Cutoff

System Requirements

VRAM requirements for different quantization methods and context sizes

1,024 tokens

58.48 GB VRAM

Consumer

3x RTX 4090

24GB VRAM

Datacenter

1x NVIDIA A100

80GB VRAM

Apple Silicon

1x Apple M3 Max

128GB VRAM

262,144 tokens

130.36 GB VRAM

Consumer

7x RTX 4090

24GB VRAM

Datacenter

2x NVIDIA A100

80GB VRAM

Apple Silicon

2x Apple M3 Max

128GB VRAM

Architecture Diagram

Evaluation Benchmarks

Rank

#55

Benchmark	Score	Rank
Agentic Coding LiveBench Agentic	0.61	9
Web Development WebDev Arena	1598	⭐ 15
Data Analysis LiveBench Data Analysis	0.77	26
General LiveBench Global	0.75	27
LiveBench Average LiveBench Average	0.75	33
Coding LiveBench Coding	0.76	42
Reasoning LiveBench Reasoning	0.80	42
Agent Arena Agent Arena	0.01	45
Mathematics LiveBench Mathematics	0.86	52
General Text Text Arena	1435	89

Rankings

Overall Rank

#55

Coding Rank

#43

About Qwen 3.8 27B

Qwen 3.8 27B is a 27-billion parameter dense foundation model by Alibaba that offers balanced efficiency and strong reasoning performance. It excels at multi-turn conversational tasks, programming, and long-context understanding.

Technical Specifications

Attention

Attention Structure

Grouped-Query Attention

Attention Heads

Key-Value Heads

Attention Head Dimension

256

Position Embedding

ROPE

RoPE Theta

10,000,000

Sliding Window Attention

Sliding Window Size

Sliding Window Ratio

Linear Attention

Yes

Linear Attention Ratio

75.0%

Normalization

RMS Normalization

Activation Function

SwigLU

Dimensions

Hidden Dimension Size

5,120

Number of Layers

FFN Intermediate Size (Dense)

17,408

Multi-Token Prediction Heads

Tokenizer

Vocabulary Size

248,320

Resources

Download Weights

About Qwen 3.8

Alibaba's Qwen 3.8 generation represents the frontier hybrid Mixture-of-Experts architecture designed for coding, professional work, research, and long-horizon agentic tasks. It features a 2.4-trillion parameter architecture (95B active per token) combining Gated DeltaNet linear attention with standard Gated Attention, available both as open weights and as a hosted flagship service.