ModelBeat
Models/Qwen3.5-4B
All models
A
qwen3-5-4b

Qwen3.5-4B is Alibaba’s open-weights AI model, released Feb 24, 2026. Its strongest use case is agentic & tools, where it ranks in the 63rd percentile of the models Model Beat tracks.

Context
Input
/ 1M
Output
/ 1M
Released
Feb 24, 2026
Modalities
Strongest atAgentic & Tools top 37%

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Qwen3.5-4B lands among all models tracked on Model Beat.

21.0
Intelligence Index
Epoch AI
21st percentile of tracked models
1.0
Coding Index
Epoch AI
1st percentile of tracked models
63.0
Agentic Index
Epoch AI
63rd percentile of tracked models

Reasoning

2 evals
GPQA Diamond77.1%

Graduate-level scientific reasoning

Humanity's Last Exam9.9%

Frontier of human expert knowledge

Coding

1 evals
SciCode16.1%

Scientific research coding

Math

1 evals
AIME 2024/202555.8%

Olympiad-qualifier math

Agentic & Tools

1 evals
τ²-bench92.1%

Tool-agent-user reliability

In the news

1

Compare Qwen3.5-4B with

Frequently asked questions

What is Qwen3.5-4B?

Qwen3.5-4B is a Qwen-family AI model, from Alibaba, released Feb 24, 2026.

Who created Qwen3.5-4B?

Qwen3.5-4B is an AI model developed by Alibaba, part of the Qwen family, released Feb 24, 2026.

How does Qwen3.5-4B perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Qwen3.5-4B scores 92.1% on τ²-bench, 77.1% on GPQA Diamond, 9.9% on Humanity's Last Exam.

Is Qwen3.5-4B open source?

Qwen3.5-4B is released with open weights (Open weights (restricted use)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Sep 4, 2026.