DataLearner logo
QW

Qwen3.8-27B

Reasoning modelCoding model

Qwen3.8-27B

Release date: 2026-08-14Updated: 2026-08-14 23:13:02.6353,738
Parameters
27B
Context length
256K
Chinese support
Supported
Reasoning ability

Qwen3.8-27B is an AI model published by Alibaba, released on 2026-08-14, for Reasoning model, with 27B parameters, and 256K context length, with a 94.60 score on MathVision.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Qwen3.8-27B

Model basics

Reasoning traces
Supported
Thinking modes
Thinking Mode (Default)Standard Mode
Context length
256K tokens
Max output length
128K tokens
Model type
Reasoning model
Modality (in / out)
Text, Image, Video → Text
Release date
2026-08-14
Model file size
No data
MoE architecture
No
Total params / Active params
27B / N/A
Knowledge cutoff
No data
Qwen3.8-27B

Open source & experience

Code license
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
GitHub link unavailable
Qwen3.8-27B

Official resources

Paper
DataLearnerAI blog
No blog post yet
Qwen3.8-27B

API details

API speed
4/5
No public API pricing yet.
Qwen3.8-27B

Benchmark Results

Qwen3.8-27B currently shows benchmark results led by LiveCodeBench (6 / 124, score 90.30), OSWorld-Verified (3 / 26, score 84.30), CharXiv RQ (2 / 15, score 90.20). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
HLE
Thinking Mode
30.80
87 / 179

Other

1 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Thinking Mode
89.20
47 / 226

Coding and Software Engineer

4 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Thinking Mode
90.30
6 / 124
SWE-Bench Pro - Public
Thinking ModeTools
61.70
10 / 57
NL2Repo-Bench
Thinking ModeTools
42.30
4 / 4
DeepSWE
Thinking ModeTools
42.20
20 / 25

AI Agent - Tool Usage

2 evaluations
Benchmark / mode
Score
Rank/total
OSWorld-Verified
Thinking ModeTools
84.30
3 / 26
Terminal-Bench 2.1
Thinking ModeTools
73
27 / 41

Agent Level Benchmark

1 evaluations
Benchmark / mode
Score
Rank/total
Agents' Last Exam
Thinking ModeTools
20.40
9 / 9

Multimodal Understanding

7 evaluations
Benchmark / mode
Score
Rank/total
MathVision
Thinking Mode
90
6 / 6
MathVision
Thinking ModeTools
94.60
2 / 6
OmniDocBench
Thinking Mode
91.10
1 / 3
CharXiv RQ
Thinking Mode
83.70
12 / 15
CharXiv RQ
Thinking ModeTools
90.20
2 / 15
BabyVision
Thinking Mode
65.70
4 / 4
BabyVision
Thinking ModeTools
85.60
2 / 4

Compare with other models

Qwen3.8-27B

Publisher

Qwen3.8-27B

Model Overview

Qwen3.8-27B is an AI model published by Alibaba, released on 2026-08-14, for Reasoning model, with 27B parameters, and 256K context length, with a 94.60 score on MathVision.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code