Open models surge (text, image, video, audio)
Key Questions
What is the latest open-weight model from Alibaba?
Alibaba released Qwen3.8-Max, an open-weight model with 2.4T parameters and multimodal support.
Which new MoE model offers large context windows?
Laguna S 2.1 is a MoE model with 118B/8B active parameters and 1M context, runnable locally in FP8.
How does Qwen Image 3.0 compare to other image generators?
It is positioned as outperforming tools like Nano Banana and ChatGPT-Image in generation quality, with APIs for editing and complex layouts.
What other open models are joining the ecosystem?
Additions include Inkling, Boogu-Image-0.1, and Audio-Visual Flamingo for diverse modalities.
Why is the open-weight ecosystem expanding rapidly?
Frontier-scale models are increasingly available for custom workflows, enabling broader access beyond closed systems.
New today: Alibaba Qwen3.8-Max open-weight model with 2.4T parameters, multimodal support; Laguna S 2.1 MoE model (118B/8B active, 1M context, local FP8). Joins Inkling, Boogu-Image-0.1, Audio-Visual Flamingo, and others. The open-weight ecosystem continues to expand with frontier-scale models becoming available for custom workflows.