Anemoi: Agent-to-Agent Coordination via Consensus-Based Planning
An agentic framework replacing centralized coordination with direct agent-to-agent communication for scalable multi-agent task solving on GAIA.
Anemoi reimagines multi-agent systems by ditching bloated centralized coordinators for a simple A2A message-passing server. Agents chat directly to align on subgoals, debate refinements, and vote on plans—slashing token waste while solving harder tasks. On GAIA, it crushes OWL-variants (+9pts) using commodity GPT-4o-mini, with 52% of gains from pure collaboration.
What changed. A2A consensus beats centralized planning by 9pts on GAIA, works with tiny planners, cuts context 37%.
Why it matters. Scales agent teams without exploding costs or needing frontier models—real path to production swarms.
Builder takeaway. Swap your orchestrator for a chatroom server; agents self-align better than you think. Paper