How should we evaluate memory for AI agents?

Log inSign up Agent Memory Leaderboard 37 posts Agent Memory Leaderboard @AgentMemoryL Agent Memory Leaderboard Official. First open championship for agent memory systems. Neutral · Reproducible · Fair. mailto: [email protected] agentmemories.ai/home Joined July 2026 67 Following 23 Followers RepliesRepliesMediaMedia Pinned Agent Memory Leaderboard @AgentMemoryL 2h The final review is underway. 🚀 We’re currently completing the final verification of code submissions and evaluation data for the Agent Memory Leaderboard. After reviewing 136+ memory systems from the community, we’re excited to share the first results on August 12. The Agent Memory Leaderboard @AgentMemoryL Aug 5 ⏳ 2 days left to submit your system to Agent Memory Leaderboard. How do we make memory evaluation fair? A leaderboard is only meaningful when different systems are compared under the same rules. That’s why AML focuses on three principles: Made with AI Agent Memory Leaderboard @AgentMemoryL Aug 3 Great to see researchers from our community sharing the Agent Memory Challenge. Building a fair and reproducible evaluation framework for agent memory requires collaboration across academia and industry. Join us and help shape the future of agent memory evaluation. 🚀 JUNDE WU @JundeMorsenWu Aug 1 We’re launching an Agent Memory Challenge. Check it out! Agent Memory Leaderboard @AgentMemoryL Aug 2 Only 5 days left for Season 1 submissions. Have an existing agent memory framework, model or API? You don’t need to build an entire evaluation stack to join the AML Leaderboard. Just 2 endpoints (Add + Search), and we handle all answer generation, scoring and benchmark Agent Memory Leaderboard @AgentMemoryL Jul 31 Every agent memory system publishes its own numbers. Almost none of them are comparable. Different datasets, different answer models, different judges. When a score moves, you can’t tell whether the memory got better or the setup got kinder. The Agent Memory Challenge has been