• •Alibaba released a preview of Qwen3.8-Max, a multimodal AI model boasting 2.4 trillion parameters.
• •The model is currently accessible via select platforms, though specific performance metrics are unavailable.
• •The absence of benchmark data leaves the model's comparative performance against competitors like GPT-4 or Claude 3.5 Sonnet unverified.

Alibaba has introduced a preview of its Qwen3.8-Max multimodal model, which reportedly features a 2.4 trillion-parameter architecture according to MarkTechPost. While Alibaba has established a track record of high-ranking open-weights models, this release deviates from standard industry practices by providing access without accompanying technical benchmarks or performance papers. The lack of standardized data creates friction for developers attempting to assess the model's capabilities in real-world reasoning tasks. Whether the model justifies its size relative to more efficient, smaller-scale competitors will remain unclear until the company releases a formal technical report or third-party evaluations emerge.