See It All. Understand It All.
Published loading...Updated

Alibaba launches and open-source Qwen3, China’s first hybrid reasoning AI model · TechNode

  • Alibaba Group launched the open-source Qwen3 large language model family on April 29, 2025, from its Hangzhou cloud division.
  • This release follows earlier Qwen2.5 models and comes amid rapid AI advancements in China despite US export restrictions on advanced semiconductors.
  • The Qwen3 lineup features eight models—including six dense and two mixture-of-experts variants—with parameters spanning hundreds of millions up to 235 billion, trained on a dataset of 36 trillion tokens and supporting 119 different languages.
  • The flagship Qwen3-235B-A22B scored 85.7 on AIME'24 and 70.7 on LiveCodeBench v5, outperforming DeepSeek R1, OpenAI o3-mini, and Google's Gemini 2.5 Pro in coding and math benchmarks.
  • Alibaba's hybrid reasoning models offer flexible thinking modes and signal increasing competition with US tech leaders, indicating sustained AI innovation in China despite restrictions.
Insights by Ground AI
Does this summary seem wrong?

19 Articles

All
Left
1
Center
2
Right
Think freely.Subscribe and get full access to Ground NewsSubscriptions start at $9.99/yearSubscribe

Bias Distribution

  • 67% of the sources are Center
67% Center
Factuality

To view factuality data please Upgrade to Premium

Ownership

To view ownership data please Upgrade to Vantage

The Tech Portal broke the news in on Tuesday, April 29, 2025.
Sources are mostly out of (0)