Alibaba Unveils 2.4-Trillion-Parameter Qwen3.8 AI Model
The announcement came during the World AI Conference in Shanghai, following recent open-weight model releases from Moonshot AI (Kimi K3) and Thinking Machines Lab (Inkling).
News
- SDAIA Unveils AI Bias Guide, Identifies Over 100 Bias Types
- AI Adoption Doubles as Saudi Internet Penetration Reaches 99.6%
- 61% of Workers Admit to Throwing a Colleague Under the Bus, Survey Finds
- Alibaba Unveils 2.4-Trillion-Parameter Qwen3.8 AI Model
- UAE Treats AI Literacy as a National Economic Priority: Report
- The Hidden Cost of Using AI at Work: It Might Be Making You Boring
[Image: Nomita Samaiyar/MITSMR Middle East]
Chinese conglomerate Alibaba has previewed its latest AI model, Qwen3.8, which it describes as “one of the most powerful models available today.”
The preview version of the model boasts 2.4 trillion parameters and is available on Alibaba’s Token Plan subscription service, alongside its Qoder and QoderWork agentic platforms, and is deemed the global second, only to Anthropic’s Fable 5.
“With massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful models available today, compatible with leading frontier AI models, second only to Fable 5,” the company said in its post on X.
The announcement was made during the World AI Conference (WAIC) in Shanghai. Earlier this year, Alibaba released the open-weight Qwen3.5 model, which has 397 billion parameters, to stay ahead in the rapidly growing market.
Qwen3.8’s announcement follows the heels of Moonshot’s unveiling of Kimi K3, a 2.8 trillion-parameter open-weight model, which it also deems “the world’s first open 3T-class model, designed for frontier intelligence.” This makes Alibaba’s product China’s second entrant in the open-weight model space, unveiled in quick succession.
Meanwhile, former OpenAI CTO Mira Murati’s Thinking Machines also launched its first open-weight AI model, Inkling, trained on 45 trillion tokens of text, images, audio, and video to serve as a “broad, balanced foundation model.”
Last month, Anthropic accused Alibaba of conducting a widespread initiative to exploit the capabilities of its AI model, Claude, in an unethical manner.
The US AI lab claimed that operators linked to Alibaba were carrying out about 29 million exchanges with its headline model by creating thousands of fake accounts. The campaign, Anthropic said, relied on “distillation attacks,” a machine-learning technique in which the output of one advanced model is used to train another.
