Stock Markets August 2, 2026 11:05 PM

Alibaba launches Qwen 3.8-MAX AI model as shares climb

Hong Kong stock rises after company unveils largest open-source Qwen model and highlights performance gains and lower inference costs

By Hana Yamamoto
Share
Twitter Reddit Facebook LinkedIn

Alibaba Group's Hong Kong-listed shares jumped after the company introduced Qwen 3.8-MAX, a Max-class update built on the Qwen 3.5 architecture that the company says improves reasoning, coding, agent and multimodal capabilities while lowering inference costs. The model, with 2.4 trillion parameters and 95 billion active parameters, will have its weights released as open source next week. Alibaba said benchmark results place Qwen 3.8-MAX broadly competitive with leading U.S. models on several coding, multimodal and engineering tests, though it trails on some general-purpose reasoning benchmarks. The move comes amid increased investment in AI infrastructure and cloud computing and heightened competition among Chinese AI firms.

Alibaba launches Qwen 3.8-MAX AI model as shares climb
Summarize with
ChatGPT Perplexity Claude Grok Gemini

Key Points

  • Alibaba introduced Qwen 3.8-MAX, an AI model built on Qwen 3.5 architecture with 2.4 trillion parameters and 95 billion active parameters.
  • The company said the model improves reasoning, coding, agent capabilities and multimodal understanding while lowering inference costs, and will release its weights as open source next week.
  • Benchmark tests reportedly show Qwen 3.8-MAX broadly competitive with leading U.S. models from OpenAI and Anthropic on several coding, multimodal and engineering benchmarks, though it trails on some general-purpose reasoning tests - sectors impacted include AI, cloud computing and technology markets.

Shares of Alibaba Group (HK:9988) climbed on Monday following the unveiling of Qwen 3.8-MAX, the companys latest flagship artificial intelligence model as domestic competition heats up in the race for advanced generative AI capabilities.

By 02:58 GMT, Alibabas Hong Kong-listed stock had risen about 6% to HK$124.00. The company said Qwen 3.8-MAX delivers enhancements across multiple dimensions of model performance, including reasoning, coding, agent functionality and multimodal understanding, while also offering lower inference costs than earlier versions.

Built on the Qwen 3.5 architecture, the new Max-class model contains 2.4 trillion parameters with 95 billion active parameters. Alibaba said it will make Qwen 3.8-MAX the first Max-class Qwen model whose weights are released as open source next week, marking a notable shift in access to its largest model weights.

According to the company, benchmark testing showed Qwen 3.8-MAX to be broadly competitive with leading U.S. AI systems from OpenAI and Anthropic. Alibaba stated the model outperformed those peers on several coding, multimodal and engineering benchmarks, while it lagged on certain general-purpose reasoning evaluations.

The launch arrives as Alibaba steps up spending on AI infrastructure and cloud services to contend with domestic rivals including DeepSeek, Baidu and Tencent, and to press on against prominent U.S. models. Earlier in May, Alibaba said it would exceed its planned AI investment of up to 380 billion yuan ($55.96 billion) over the next three years.


Context and competitive dynamics

The rollout follows a period of rapid model upgrades across Chinese AI companies. Alibaba highlighted that competition accelerated after DeepSeeks introduction of low-cost reasoning models, which reshaped the competitive landscape earlier this year.

Alibaba emphasized lower inference costs for Qwen 3.8-MAX compared with prior iterations, and pointed to benchmark advantages in coding, multimodal and engineering tasks versus some leading U.S. models, balanced by weaker outcomes on some general-purpose reasoning tests.


Bottom line

Qwen 3.8-MAX represents a significant technical update by Alibaba, accompanied by a public release of model weights and an asserted improvement in cost efficiency for inference. The announcement coincided with a marked positive move in Alibabas Hong Kong-listed shares and underscores intensifying competition among major AI developers in China and abroad.

Risks

  • Qwen 3.8-MAX trails leading U.S. models on some general-purpose reasoning tests, which may limit its competitive standing on certain tasks - this affects AI model adoption and applications relying on general reasoning.
  • Competition in Chinese AI intensified after DeepSeeks low-cost reasoning models reshaped the landscape earlier this year, creating a faster upgrade cycle and greater execution risk for firms in the AI and cloud sectors.
  • Alibaba plans to exceed its planned AI investment of up to 380 billion yuan over the next three years, introducing financial commitment and execution risk in its AI infrastructure and cloud computing strategy.

More from Stock Markets

Nvidia Retreats From 2026 Peak - Does the Pullback Create a Buying Opportunity? Aug 2, 2026 FleetPartners Draws A$760.3 Million Bid From Pacific Equity Partners, Shares Jump Aug 2, 2026 KOSPI Slides Over 4% as Chip Leaders Drag Market Following July Rout Aug 2, 2026 Steadfast Shares Rise After U.S.-Led Group Reaffirms A$6.00-Per-Share Takeover Proposal Aug 2, 2026 Capital One Tells Court Account Closures Followed Anti-Money Laundering Review Aug 2, 2026