Qwen3

Think Deeper or Act Faster

5.0•14 reviews•

1.9K followers

Think Deeper or Act Faster

5.0•14 reviews•

1.9K followers

Visit website

LLMs

Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud. - QwenLM/Qwen3

Have a question about Qwen3? Ask it here and get a real answer.

AppSignal — Full-stack monitoring for errors, metrics, and logs

Full-stack monitoring for errors, metrics, and logs

Promoted

Qwen3 Launches

Qwen3.5 Small 0.8B-9B native multimodal w/ more intelligence, less compute

Launched on March 3rd, 2026

Qwen3.5 The 397B native multimodal agent with 17B active params

Launched on February 17th, 2026

Qwen3-TTS Voice design, cloning & 97ms streaming

Launched on January 23rd, 2026

Qwen-Image-2512 SOTA open-source T2I model with even greater realism

Launched on January 1st, 2026

Qwen-Image-Layered Turn flat images into multi-layer editable assets

Launched on December 22nd, 2025

View all Qwen3 launches

Forum Threads

p/qwen3

•

2mo ago

What a year for Qwen!

Check out the epic Super Qwenrio adventure in 2025. Really looking forward to 2026!

View all

Qwen3

Think Deeper or Act Faster

Think Deeper or Act Faster

Have a question about Qwen3? Ask it here and get a real answer.

Qwen3 Launches

Forum Threads

What a year for Qwen!

Have a question about Qwen3? Ask it here and get a real answer.

Qwen3 Launches

Forum Threads

What a year for Qwen!

What's great

vs Alternatives

Is memory usage manageable at 1M context with vLLM?

How robust is alignment and refusal behavior across languages?

Were the recommended versions necessary to avoid bugs?

What's great

What needs improvement

vs Alternatives

How accurate is Qwen3 on real coding tasks you tried?

Does Qwen3-Coder reduce PR review time or defects?

Alternatives Considered

What's great

vs Alternatives

Is memory usage manageable at 1M context with vLLM?

How robust is alignment and refusal behavior across languages?

Were the recommended versions necessary to avoid bugs?

What's great

What needs improvement

vs Alternatives

How accurate is Qwen3 on real coding tasks you tried?

Does Qwen3-Coder reduce PR review time or defects?

Alternatives Considered