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
an hour ago
- South China Morning Post
Talks deadline for CK Hutchison US$23 billion Panama ports deal passes on Sunday
A deadline for exclusive talks on a US$23 billion sale of global port stakes by Hong Kong tycoon Li Ka-shing's CK Hutchison Holdings passed without a deal on Sunday, with analysts expecting complex negotiations to be extended amid intense US-China geopolitical rivalry. Advertisement The controversial transaction involved CK Hutchison selling stakes in 43 ports, including two at either end of the Panama Canal, to a consortium led by Terminal Investment Limited, an affiliate of the world's largest container line, MSC, and American asset manager BlackRock. The July 27 deadline was set 145 days from the company's March 4 exchange filing that first announced the exclusive negotiation period. Shipping and legal experts earlier told the Post that they were not optimistic the deal would be signed in its original form by Sunday, saying it could be subject to substantial changes given the political headwinds and regulatory hurdles in both Panama and mainland China. By midnight on Sunday, Hutchison had not disclosed any information on the deal. Amid the ongoing trade tensions, a high-level American business delegation was expected to visit Beijing this week, the Post reported exclusively. Sources said the trip would be organised by the US-China Business Council.
South China Morning Post
2 hours ago
- South China Morning Post
China, US to extend tariff pause at Sweden talks by another 90 days: sources
Beijing and Washington are expected to extend their tariff truce by another three months at trade talks in Stockholm beginning on Monday, according to sources close to the matter on both sides. China and the United States agreed in May to remove most of the heavy tariffs levied on each other's goods for 90 days while continuing trade negotiations. That suspension is set to expire on August 12. During the third round of trade negotiations between the world's two biggest economies, both will expound their views on major sticking points – such as the US' concerns over China's industrial overcapacity – rather than achieve specific breakthroughs, the sources said. One source said that, during the expected 90-day extension, the two nations will commit to not impose additional tariffs on each other, nor escalate the trade war by other means. According to three people familiar with Beijing's position, while the earlier discussions in Geneva and London focused on 'de-escalation', in the latest meeting the Chinese delegation will also press Trump's trade team on fentanyl-related tariffs.
South China Morning Post
2 hours ago
- South China Morning Post
‘Global approach' to AI regulation urgently needed, UN tech chief says
The world urgently needs to find a global approach to regulating artificial intelligence, the United Nations' top tech chief said this week, warning that fragmentation could deepen risks and inequalities. Doreen Bogdan-Martin, head of the UN's International Telecommunications Union (ITU) agency, said she hoped that AI 'can actually benefit humanity.' But as concerns mount over the risks posed by the fast-moving technology – including fears of mass job losses, the spread of deepfakes and disinformation, and society's fabric fraying – she insisted that regulation was key. 'There's an urgency to try to get … the right framework in place,' she said, stressing the need for 'a global approach.' Her comments came after US President Donald Trump this week unveiled an aggressive, low-regulation strategy aimed at ensuring the United States stays ahead of China on AI. Among more than 90 proposals, Trump's plan calls for sweeping deregulation, with the administration promising to 'remove red tape and onerous regulation' that could hinder private sector AI development.
