Sage-A1.5 is a 35B MoE model for long-horizon agents
Original title: Sage-A1.5: From Long-Horizon Execution to Scientific Exploration with Capability Composition
It adds subagent delegation, progress checks with strategy revision, and exploratory research beyond predefined questions.
The report introduces Sage-A1.5, a 35B mixture-of-experts agentic model built on A1 for long-horizon problem solving through adaptive capability composition. Orchestration uses task delegation and parallel subagents with isolated contexts. Verification adds progress checking, evidence validation, and reflective strategy revision. The authors also describe open scientific discovery that produces inspectable research artifacts and say they hope it becomes an accessible community foundation.
Unconfirmed: The report does not state whether Sage-A1.5 weights are publicly released.
- Sage-A1.5 is a 35B mixture-of-experts model that builds on A1.
- Subagents work concurrently with isolated contexts to limit interference.
- Verification is meant to abandon unproductive search paths and try alternatives.
- Capabilities are shown on complicated agentic tasks and multidisciplinary scientific applications.
From the original article
1 min read at huggingface.co
In 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, Tianshuo Peng, Shiyang Feng · huggingface.co
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.