Sage-A1.5 is a 35B MoE agent for long-horizon scientific work
Original title: Sage-A1.5: From Long-Horizon Execution to Scientific Exploration with Capability Composition
The report says the model delegates work to subagents, checks its own progress, and produces inspectable research artifacts.
A technical report introduces Sage-A1.5, a 35B Mixture-of-Experts agentic model that builds on A1. It adds task delegation and parallel subagents with isolated contexts, plus progress checking, evidence validation, and strategy revision. The authors also describe exploratory scientific work that yields interactive analysis and inspectable artifacts, and say they want it to be an accessible foundation.
Unconfirmed: The report does not state whether model weights are publicly released.
- Sage-A1.5 is a 35B Mixture-of-Experts model built on A1.
- Subagents run concurrently with isolated contexts to limit interference.
- Verification covers progress checks, evidence validation, and strategy revision.
- Demonstrations cover complex agent tasks and multidisciplinary science.
From the original article
1 min read at arxiv.orgIn this report, we introduce Sage-A1. 5, a 35B Mixture-of-Experts agentic model that builds on A1 to advance long-horizon problem solving through adaptive capability composition. Compared with A1, Sage-A1.
By Wenjie Lou · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.