cd /news/ai-agents/agent-memory-benchmarks-need-more-th… · home › topics › ai-agents › article
[ARTICLE · art-173155] src=x.com ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

Agent memory benchmarks need more than an overall score

The Agent Memory Challenge (AML) Cycle 2 results drew 402 task submissions across text, code and multimodal memory tracks and generated more than 1 million cross-platform impressions and over 10 million cumulative website clicks, according to the @AgentMemoryL leaderboard organizers. Khipu, a system that stores every message verbatim and converts photos to text for search, took 1st place on the Multimodal Open-Source leaderboard with 43.26 points, while Gedo Memory placed 3rd on the Text Memory Commercial Products board and 5th among closed-source methods. Analysis from @omarsar0 showed capability breakdowns that overall rankings can hide: on the closed-source text board the top 10 systems scored between 70 and 78 on Memory Governance, and Personalization & Care was the lowest-scoring category.

recommendation

Break agent memory evals into per-capability categories like Memory Governance and Personalization & Care instead of relying on one overall score, since top-10 systems clustered within 8 points on governance.

more in playbook: Evals and testing for LLM apps
by read1 min views8 publishedNov 20, 2026
Agent memory benchmarks need more than an overall score
Image: source

🏆 AML Cycle 2 results are live! 💥 • 1M+ cross-platform impressions • 10M+ cumulative website clicks • 402 task submissions • Three tracks: text, code and multimodal memory This cycle brought together teams from across the global AI community, including teams fromShow more

A closer look at AML Cycle 2 from @omarsar0: capability breakdowns reveal differences that overall rankings can hide. Personalization & Care stands out as an area for further investigation. Thanks for the thoughtful analysis—we’ll share more detailed breakdowns and cases soon!

I like how this measures different aspects of agent memory capabilities. This is part of the @AgentMemoryL Cycle 2 results. On the closed-source text board, the top 10 systems score between 70 and 78 on Memory Governance. Personalization & Care is the lowest-scoring category

Khipu took 1st place in the AML Cycle 2 Multimodal Open-Source leaderboard with 43.26 points! 🥇 Congratulations to @Lv999APTITUDE, and thank you for contributing to the new track. Excited to see what comes next for the project!

Khipu: memory for AI agents. Stores every message verbatim, never rewrites it; photos become text for search and come back as evidence. 1st in Multimodal · Open-source Methods, Agent Memory Challenge Cycle 2. @AgentMemoryL#AgentMemoryLeaderboardgithub.com/Lv999APTITUDE/…

Congratulations to @gedolabs on Gedo Memory’s 3rd-place finish in the AML Cycle 2 Text Memory Commercial Products leaderboard! 🥉 Great to see the team sharing lessons from the evaluation and putting them into practice. Looking forward to what comes next👏

Gedo Memory(gedo.ai) placed 3rd on the AML Cycle 2 Commercial Products leaderboard (Text Memory), and 5th among closed-source methods. I'm grateful to our team, and to the AML and CSIG organizers. A few notes on what this meant for us. Why we entered At Gedo

── more in #ai-agents 4 stories · sorted by recency
── more on @agent memory challenge 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/agent-memory-benchma…] indexed:0 read:1min 2026-11-20 · —