Alibaba Group Holding continues to scale up its artificial intelligence efforts with the release of its biggest AI model to date, joining the ranks of OpenAI, Google DeepMind and other major players with trillion-parameter models.
Advertisement
Qwen-3-Max-Preview, the company’s first model to have more than 1 trillion parameters, was released on Alibaba’s official cloud-services platform and the large language model marketplace OpenRouter on Friday.
The model is the latest in the Hangzhou-based company’s Qwen3 series, first released in May, which initially consisted of models ranging from 600 million to 235 billion parameters. Alibaba owns the South China Morning Post.
Parameters are the variables that encode an AI system’s “intelligence” and are adjusted during the training process. Generally, a higher number of parameters means stronger capabilities, although it also means more computational power is needed to train and run the model.
OpenAI’s GPT-4.5, widely believed to be one of the industry’s biggest models, has an estimated parameter count of 5 to 7 trillion.
Advertisement
According to Alibaba, Qwen-3-Max-Preview is a text-only model that outperformed its previous “best” model, the Qwen3-235B-A22B-2507 released in July, on internal tests.
It also published scores suggesting that the model had bested MoonShot AI’s Kimi K2, a non-reasoning version of Anthropic’s Claude Opus 4 and DeepSeek V3.1 on five benchmarks, although these were not released as part of an official technical report for the model.