SciMARS: Multi-Agent Recursive Self-Improvement for Closed-Loop Scientific Discovery
- Published
- Source
- arXiv
- Paper number
- 1420
- Field
- Agents
- arXiv ID
- 2611.12552
Key points
- SciMARS is a recursive self-improvement framework where separate editors independently revise role procedures and coordination contracts of a multi-agent team.
- A shared evidence graph records observations, findings, plans, and unresolved questions with provenance and routes them across roles.
- SciMARS + Memory adds a third editor for how agents record, retrieve, and receive evidence.
- On Autonomous Science Bench with GPT-6-Astra, it resolved 12/18 attempts, ahead of 8/18 for the base team.
- It solves 14/38 ScienceAgentBench tasks and 11/20 AppWorld tasks, compared with 11/38 and 5/20 for the base team.
- In the GLM-5.3 cohort, it resolved 2/18 attempts, slightly ahead of 1/18 for the base team.
Paper links
External research summaries. These are not HDATF publications or measured product results.