Stanford and Together AI paper shows self-organizing agent teams beat routing
The non-obvious finding is that letting agents write their own teamwork protocol beats even an oracle that picks the best individual answer, and the protocol transfers across benchmarks after training on just 15 problems. Builders of multi-agent systems can copy this approach instead of hand-tuning roles and pipelines, with off-the-shelf models like DeepSeek-V3 already sufficient as team members.
A paper from Stanford and Together AI shared by DAIR.AI shows self-organizing multi-agent teams outperform individual models and router selection. A team using o3-mini, Claude Sonnet 4 and DeepSeek-V3 averaged 66.7% across five math and physics benchmarks, versus 48.8% for its strongest member alone and 59.0% for a perfect router. On AIME 2026 the team scored 71.2%, 13.4 points above the router. One agent reviews past exchanges and rewrites the team's collaboration strategy, covering roles, discussion order and answer merging. The strategy was learned from only 15 AIME 2024 problems and generalized unchanged to held-out problems and four new benchmarks.
Recommended paper. Research on effective agent collaboration/communication is lacking, but self-organizing agents seem much more capable than previously thought. DAIR.AI @dair_ai Banger paper from Stanford and Together AI. They show why it might be a good idea to let your agent team learn its own way of working together. (bookmark it) Three models (o3-mini, Claude Sonnet 4 and DeepSeek-V3) averaged 66.7% across five math and physics benchmarks as a self-organizing team. Their strongest member alone scored 48.8%, and a perfect router choosing among the members' independent answers scored 59.0%. On AIME 2026 the team reached 71.2%, 13.4 points above that router. One member reviews the team's earlier exchanges and rewrites the teamwork strategy, covering roles, the order of discussion phases, who participates and how partial answers are combined. The strategies were learned from only 15 AIME 2024 problems and then applied unchanged to held-out AIME 2024 problems and four new benchmarks. Paper: academy.dair.ai/papers/self-or… 🔗 View Quoted Tweet 💬 4 🔄 0 ❤️ 5 👀 1569 📊 4 ⚡
