IreneAI
AML Cycle 2: What to look for when the results drop
Agent Memory benchmarks should measure task success, not retrieval alone
Can Coding Agents Reuse Past Debugging Experience Without Repeating Mistakes?
Evaluating Memory Systems for AI Agents Is Harder Than It Looks
Show HN: AML CYCLE 2, a benchmark for long-term agent memory
Beyond Context Windows: Evaluating Long-Term Memory for AI Agents
Evaluating Long-Term Memory for AI Agents: AML Cycle 2 Is Now Open
How should we evaluate whether an AI agent's memory is still current?
Show HN: An open Add/Search evaluation framework for agent memory
https://agentmemoryleaderboard.ai/
Hi HN,We are trying to make different Agent Memory systems comparable without letting each team choose its own answer model and evaluation pipeline.
Coding Memory Helps Agents Reuse Engineering Experience
GraphMemix: Query-Aware Evidence Forests for Long-Term Multimodal Agent Memory
What should a fair benchmark for Agent Memory look like?
Show HN: Agent Memory Leaderboard – first public results for AI memory systems
https://agentmemoryleaderboard.ai/leaderboard/academic/textual
We just released the first public results from the Agent Memory Leaderboard (AML).The first evaluation focuses on Text Memory across two tracks:- Open-source Methods - Commercial Products136 teams registered, and 69 representative memory frameworks completed the first evaluation.Commercial Products — Text Memory:1. MemoraX — 58.02 2. MemOS — 45.89 3. NTES-MEMORY-SMART — 44.21The benchmark uses...
Re:How should we evaluate memory for AI agents?
on Tuesday August 11, 2026 @10:57PM (#49267348)
> Sorry for the concern, and thank you for your support of AML. Your submission is eligible.
>
> According to the AML requirements, if an LLM is used, it must be *GPT-4o-mini*. If no LLM is used, this requirement does not apply.