Meridiano

Add Article

No Image Available

Article Details

Title

Article: Alibaba unveils Qwen-3.5, sharpening global race to spread AI models

Impact Score

6 / 10

AI Summary (Processed Content)

Alibaba Cloud has released its next-generation open AI model, Qwen-3.5, intensifying the global AI competition between China and the U.S. The launch was part of a wave of new model releases from major Chinese AI developers.

The new open-source model shows significant performance improvements over its predecessor and is benchmarked as competitive with leading U.S. models from firms like OpenAI. A separate closed-source version, Qwen-3.5-Plus, boasts a massive context window and performance claimed to be on par with current industry leaders.

The main topics covered are Alibaba's Qwen-3.5 AI model release, the China-U.S. AI race, and the technical specifications and benchmarked performance of the new models.

Original URL

https://www.scmp.com/tech/big-tech/article/3343738/alibaba-unveils-qwen-35-sharpening-global-race-spread-ai-models?utm_source=rss_feed

Source Feed

Big Tech - South China Morning Post

Published Date

2026-02-16 09:19

Fetched Date

2026-03-04 14:30

Processed Date

2026-03-04 15:21

Embedding Status

Present

Cluster ID

Not Clustered

Raw Extracted Content

Alibaba unveils Qwen-3.5, sharpening global race to spread AI models
With multimodal capabilities and open weights, Qwen-3.5 signals Alibaba’s ambition to anchor the next phase of global AI deployment
Alibaba Cloud has unveiled its next-generation open artificial intelligence model, Qwen-3.5, in a move set to have major implications for the global AI race between China and the US.
The much-anticipated release, timed on the eve of the Lunar New Year, capped a frenetic week in which nearly every major Chinese AI developer rolled out new flagship models.
The two new 3.5-series models were made available on Alibaba’s cloud platform Model Studio on Monday afternoon.
The new Qwen-3.5-Open-Source model with 397 billion parameters showed significant improvement over the company’s previous flagship model, Qwen-3-Max-Thinking, despite the latter’s much larger size of over 1 trillion parameters, according to self-reported benchmark scores.
The model also boasted a performance on par with leading models from OpenAI, Anthropic and Google DeepMind, the benchmarks showed, although the comparison was not with the three US heavyweights’ latest models.
Meanwhile, a closed-source version called Qwen-3.5-Plus achieved performance “on par with state-of-the-art leading models” and has a listed context window – the amount of data it can process at any given time – of 1 million tokens, one of the largest in the industry.