Hi ,
We wanted to invite you to join AgentX–AgentBeats, a competition hosted by Berkeley RDI in conjunction with the Agentic AI MOOC and its global community of ~40K registered learners. AgentX–AgentBeats aims to advance the state of the art in agentic AI by creating benchmarks, and then AI agents that top those benchmarks.
The competition runs in two phases:
Phase 1 (Completed)
Participants build green (or evaluator) agents that define assessments and automate scoring. A green agent provides a specific agent evaluation benchmark including the environment, a set of tasks, and the evaluator.
More than 3,000 participants across 1,300 teams joined Phase 1, and we were thrilled to see all of the participants’ creativework!
Phase 2 (Now Live)
Participants build “purple agents” that attempt to perform well on Phase 1 agentified benchmarks, competing on public leaderboards.
Phase 2 involves a sprint-based format, with four rotating sprints over the next couple of months:
🗓️ Sprints Schedule
- Sprint 1 (3/2 - 3/22): Game Agent, Finance Agent, Business Process Agent
Sprint 2 (3/23 – 4/12): Research Agent, Multi-agent Evaluation, τ²-Bench, Computer Use & Web Agent
Sprint 3 (4/13 – 5/3): Agent Safety, Coding Agent, Cybersecurity Agent
Sprint 4 (5/4-5/24): General Purpose Agents, the grand finale of AgentBeats Phase 2, where everything culminates. AgentX–AgentBeats is the first competition to explicitly spotlight general-purpose agents, testing broad capability, adaptability, and robustness across diverse tasks rather than a single domain. While earlier sprints emphasize depth, this final sprint showcases breadth and real-world readiness.
For more detailed info about all sprints, we encourage you to review the AgentX–AgentBeats website!
