Alibaba upgrades flagship Qwen3 model to outperform OpenAI, DeepSeek in maths, coding
The new Qwen3-235B-A22B-Instruct-2507-FP8 is an open-source model that achieved 'significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage', according to a Tuesday update on artificial intelligence (AI) community HuggingFace and ModelScope, Alibaba's open-source platform. Alibaba owns the Post.
It outperformed some rivals in certain assessments, such as the 2025 American Invitational Mathematics Examination, where the new Alibaba model scored 70.3. By comparison, DeepSeek-V3-0324, the most recent version of the foundational model that was released in March, scored 46.6 while OpenAI's GPT-4o-0327 scored 26.7.
As for coding capabilities, the new Qwen secured 87.9 points from the MultiPL-E benchmark, slightly higher than 82.2 and 82.7 from the DeepSeek and OpenAI models above, respectively, though it lagged behind Claude Opus 4 Non-thinking, from Anthropic, which scored 88.5.
Alibaba's new release was an upgrade from the Qwen3-235B-A22B-FP8. But it only supports non-thinking mode, where an AI system provides a direct output without the explicit reasoning steps or chain of thought that a thinking model might employ. As a result, its content length was boosted eightfold to 256,000 tokens, making it able to handle longer texts in a single conversation.
Also on Tuesday, Alibaba said a Qwen model with 3 billion parameters would be integrated into HP's smart assistant 'Xiaowei Hui' on its personal computers in China, enhancing capabilities including drafting documents and summarising meetings.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
South China Morning Post
5 hours ago
- South China Morning Post
China-EU relations at 50
China-EU relations at 50 This year marks half a century of formal diplomatic relations between China and the European Union as well as the 25th anniversary of the founding of the European Union Chamber of Commerce in China. In this series, the Post examines the ties between the two powers.
South China Morning Post
7 hours ago
- South China Morning Post
‘Rebalancing' needed in China-Europe relationship, chamber president says
This year marks half a century of formal diplomatic relations between China and the European Union, as well as the 25th anniversary of the founding of the European Union Chamber of Commerce in China. In this entry of our series examining ties between the two powers, Ji Siqi speaks to the chamber's president about business sentiment in a tense period for global trade. Advertisement The president of the European Union Chamber of Commerce in China – the chief non-profit organisation advocating on behalf of the continent's businesses – has said the relationship between Beijing and Brussels has reached a tipping point, encouraging the two to realign their collaborative model and distribute benefits in a more equitable manner. Jens Eskelund said there is a strong perception among the European population that China is taking most of the spoils from bilateral trade, as the EU's manufacturing sector struggles to compete with a glut of cheaper goods. 'When we look back at the past 50 years of the bilateral relationship, it has created enormous value for both sides,' Eskelund told the Post on the eve of the chamber's 25th anniversary. 'Chinese exports have created jobs and wealth in China, and given the average European higher purchasing power. 'Now the question is, if we are in a situation where very intense pressure from China leads to losses for European companies … then, of course it becomes, 'Hang on, why are we doing this?'' Advertisement The relationship between China and the EU has been fraught in recent years, despite continuous dialogue as both sides seek to avoid the sort of full-blown trade war being waged by US President Donald Trump.
South China Morning Post
8 hours ago
- South China Morning Post
Chinese scientists break design ‘curse' that killed US Navy's X-47B drone programme
Chinese aerospace engineers have a revolutionary software design, which they say will allow them to overcome a major barrier to stealth aircraft development The new platform allows plane designers to have as many design variables as they want without increasing computing load – a feat long deemed impossible in aviation circles. The researchers described their innovation as breaking the 'dimensionality curse' and used the US Navy's X-47B, a demonstration stealth drone, to illustrate how the system worked. Once celebrated for its carrier landings and autonomous aerial refuelling, the X-47B project was cancelled in 2015 because of unresolved trade-offs between stealth, aerodynamics and propulsion. However, the Chinese software design delivered dramatic improvements to the design with 740 variables, including measures to reduce flight drag and its radar signature, as well as improving engine thrust while maintaining airflow stability. 'Traditional global optimisation algorithms face the curse of dimensionality problem,' wrote the team led by Huang Jiangtao from the China Aerodynamics Research and Development Centre in a peer-reviewed paper published in Acta Aeronautica et Astronautica Sinica earlier this month. The shape of components such as wing leading edges and engine inlet ducts affects two crucial things: how smoothly the plane flies and how easily it can be detected by enemy radars.
