Portuguese 
Spanish 
Alibaba Unveils Qwen-3-Max-Preview, AI Model with 1 Trillion Parameters, Billion-Dollar Investments, and Focus on Competing with OpenAI and Google.
Alibaba has introduced its most advanced artificial intelligence model, the Qwen-3-Max-Preview, with over 1 trillion parameters. The announcement took place on Friday, on its official cloud platform and on OpenRouter, one of the leading global language model marketplaces.
This model is the evolution of the Qwen3 series, launched in May. Initially, the line featured versions ranging from 600 million to 235 billion parameters. Now, the Chinese giant is betting on an unprecedented scale to face international competitors.
Direct Competition
Parameters define the accuracy of an AI model. However, the larger the quantity, the greater the energy consumption and computing power required.
-
ChatGPT Alerts FBI to Prevent Planned Murder of 8-Year-Old in Brazil’s Espírito Santo
-
Scientists Discover 125-Million-Year-Old “Pregnant Fossil” with Preserved Embryos and Microscopic Larvae Inside Mollusk in England
-
17-Year-Old Justin Bernstein Develops Bacteria to Produce Glass on Arctic Ice to Reflect Sunlight, Wins $50,000 U.S. Scholarship
-
The 19-Meter Rocket-Helicopter: A Private Industry Prototype That Challenged Space Race Logic
The OpenAI GPT-4.5 is estimated to have between 5 and 7 trillion parameters, holding a prominent position in the market. With the Qwen-3-Max-Preview, Alibaba surpasses its previous model, the Qwen3-235B-A22B-2507, launched in July.
The company aims to position itself as a rival not only to OpenAI but also to Google and other major developers.
Released Tests
Alibaba stated that the Qwen-3-Max-Preview surpassed the Kimi K2 from MoonShot AI, a non-rational version of Claude Opus 4 from Anthropic, and DeepSeek V3.1 in five benchmarks.
However, an official technical report did not confirm the results. Nonetheless, the company highlighted significant advancements in understanding texts in Chinese and English, executing complex instructions, and multilingual capabilities.
Highlighted Advances
According to the company, the model showed improvements in performing open tasks, using tools, and scalability.
Engineer Binyuan Hui stated that a “thinking” version is in development. This update could further expand the potential of the series.
Impact on the Ecosystem
The Qwen models have already secured Alibaba’s leadership in the global open-source AI ecosystem, with over 20 million downloads and 100,000 derived models on the Hugging Face platform.
Despite this, the Qwen-3-Max-Preview has not yet been made available as open-source, similar to what happened with the Qwen2.5-Max. Access, for now, is only through official channels.
Announced Prices
The new model has tiered pricing: US$ 0.861 per million input tokens and US$ 3.441 per million output tokens. It is one of the most expensive in the Qwen series.
In comparison, the Qwen3-235B charges US$ 0.287 per million input tokens and US$ 1.147 per million output tokens. Meanwhile, the Kimi K2 charges US$ 0.60 and US$ 2.50, respectively.
Billion-Dollar Investment
To support this strategy, Alibaba announced an investment of 380 billion yuan (US$ 52 billion) in AI infrastructure over the next three years.
This amount surpasses the company’s spending on artificial intelligence over the last decade. Therefore, the move clearly indicates the intent to dominate the sector.
The Role of China
Alibaba is among the largest Chinese investors in AI. In addition to advanced models, reports indicate that the company is developing a national processor to reduce dependence on Nvidia chips, which are manufactured in the United States.
This initiative follows Beijing’s guidelines, which have tightened control over imports of foreign semiconductors.
Furthermore, it reinforces the technological competition between China and the U.S., with direct reflections in the global artificial intelligence market.
Strategic Consolidation
The launch of the Qwen-3-Max-Preview marks another step for the giant from Hangzhou to consolidate its position among global leaders.
The model contributes to the international race around large-scale AI and shows that China wants to actively participate in the next phase of this technology.
