✦
Summary
Sage-A1.5 is a 35B mixture-of-experts agent model for engineering, scientific research, and long-horizon information work. It extends Agents-A1 with capability composition, independent-context orchestration, and an optional internal verification mode.
About Sage-A1.5
Developers can study sustained agent execution and scientific problem solving with public model distributions and project resources. Core capabilities include 256K served context, Agentic task orchestration, Opt-in self-review, and Software engineering workflows.
Public code or model weights are linked from the project, with setup instructions for running or adapting the system. Additional characteristics include Long-horizon search, Scientific problem solving, Instruction following, and Tool-calling workflows.