Agent Memory Leaderboard - Cycle 1 - Academic: leaderboard
Agent Memory Leaderboard (2026): cycle-one textual public_suite_v3, 0–100. Academic and industry tracks remain separate. Rows identify complete memory systems and their submitted versions, not their underlying language model; aggregate average_score is retained at source precision. Score (%).
Metric: Score (%). Source: agentmemoryleaderboard.ai. Saturation forecast: Rough model projection: around 2026. 50 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | AML InvMem | 45.06 | |
| 2 | AML Refind | 44.97 | |
| 3 | AML ActiveMemoryIndex | 44.84 | |
| 4 | AML Hybrid Search v2.0 | 44.57 | |
| 5 | AML ChronoHybridMem | 44.33 | |
| 6 | AML Hybrid Episodic Memory | 44.28 | |
| 7 | AML Chronicle | 44.19 | |
| 8 | AML FlowGrid_AML_Retriever | 43.98 | |
| 9 | AML LLLMemoryAgent | 43.06 | |
| 10 | AML aml-memory-mvp | 42.95 |
Interactive version: theaggregate.ai/benchmark?slug=agent-memory-leaderboard-cycle-1-academic · How It Works · Data refreshed daily, snapshot 2026-10-21.