China’s Alibaba (9988.HK), on Monday introduced its largest and most capable AI model to date, sending its shares increasing, while a research firm stated that DeepSeek’s latest product provide cut-throat pricing that is more than 100 times cheaper than Anthropic’s Claude Fable 5.
The two development emphasize the fast pace of development in artificial intelligence by Chinese tech firms, which are locked in a fierce and fast-moving struggle to construct more powerful systems without making them prohibitively costly to run.
Both models — Alibaba’s Qwen3.8-Max and DeepSeek’s V4-Flash — underline Chinese dedication to open-weight models as the firms seek to gain traction amongst developers globally.
“Chinese AI companies have discovered an vital market. Many business workflows do not require the industry’s very best model,” stated Lian Jye Su, chief analyst at research firm Omdia.
“They want models which are good enough, budget friendly, transparent and accessible, and open-weight models support meet that demand.”
With an open-weight model, the underlying learned settings that permit developers to run or adapt the system are available for download. By evaluation, OpenAI, Anthropic and Google (GOOGL.O), have closed-source models.
Trillions Of Parameters
Alibaba’s new Qwen3.8-Max right away shot up leaderboards assessing the capabilities of AI models after being introduced on Monday, supporting its shares climb 7% in Hong Kong trade.
The model has 2.4 trillion parameters, the numerical settings a model learns from data and uses to recognize patterns, generate answers, and carry out tasks. That puts it not far behind domestic rival Moonshot AI’s Kimi K3, which was released last month and has 2.8 trillion parameters.
A higher parameter figure does not automatically make a model better, but it has emerge as a closely watched measure of the scale of the computing and data behind advanced AI systems.
Qwen3.8-Max was introduced on crowdsourced, model-comparison platform Arena.AI. It quickly became the highest-ranking Chinese model in terms of text models, although it still lags Claude Fable 5 and 3 Opus variants that are all from Anthropic.
On Arena.AI’s leaderboard for AI models that examine images and other visual material, Qwen3.8-Max ranked second worldwide, only behind a Claude Fable 5 variant.
Both Qwen3.Eight-Max and Kimi K3 can handle text, images and video, and process up to 1 million tokens at a time.
Tokens are chunks of data, frequently parts of words or short words, and a huge figure means the model can take in huge amounts of material in one go, together with long legal files, a large software codebase or hundreds of pages of documents.
The tech giant said the model, because of be launched next week, finished a software-engineering project in 16 days. It uses a “combination-of-experts” design, which divides work among specialized parts of the system rather than of switching on the whole model for every request. Only 95 billion parameters are used at a time, cutting costs and response delays.
