Agent Evaluation & Benchmarking Summit

Evaluating agents is one of the hardest problems in the space. This one-day summit brings together the teams building evaluation frameworks — from task-specific benchmarks to end-to-end reliability metrics.

Hear from the creators of SWE-bench, GAIA, and AgentBench on what they've learned. Workshop sessions cover building custom evals, A/B testing agent behaviors, and measuring safety properties. If you're shipping agents to production, this is how you know they work.