LOGOS-8B
It supports six downstream tasks: interaction-aware ligand design for binding pockets, protein ligand-binding site identification, retrosynthesis prediction, unconditional material generation, protein editing, and antibody CDR design. The unified grammar aligns pre-training and downstream objectives, allowing strong cross-domain generalization. LOGOS is trained at 1B, 3B, and 8B scales with consistent positive scaling. The 8B variant is built on a Qwen3 architecture with continued multi-domain pre-training. Developed by Alibaba Group and the Gaoling School of AI, Renmin University of China. Weights released under CC BY 4.0.
Overview
LOGOS-8B (Language Of Generative Objects in Science) is an 8B parameter autoregressive scientific foundation model. It encodes proteins, antibodies, small molecules, chemical reactions, and materials into a unified scientific grammar, enabling a single model to perform generation, prediction, and design tasks across the natural sciences without relying on 3D geometry.
About Alibaba
Chinese e-commerce and cloud leader behind Taobao, Tmall, and Alipay.
View Company ProfileTools using LOGOS-8B
No tools found for this model yet.
MongoDB - Build AI That Scales
