{"slug": "enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well", "title": "Enki – memory for AI agents that keeps ~half as much and answers as well", "summary": "Enki Labs released evaluation results for its closed-source memory engine Enki, showing comparable answer accuracy to mem0 on the LongMemEval-S benchmark while storing roughly half the facts (138 vs 283). In a 25-instance validated slice, Enki achieved 14/25 correct answers versus mem0's 12/25, with a notable advantage in multi-session reasoning (4/5 vs 2/5).", "body_md": "Enki is a memory engine for LLM agents. **This repository publishes evaluation results only** — the engine is closed-source. No configuration, internals, or methodology beyond what is described below is included here.\n\nBoth systems ingest **identical** conversation histories from LongMemEval-S. Each system's\nretrieved memories are answered by the **same** model (Claude Haiku) and graded by the\n**same** LLM-as-judge, at equal retrieval depth (K=10). The only variable is the memory layer.\n\n**Validated slice: 25 instances** (full-benchmark run in progress).\n\n| Question type | Enki | mem0 |\n|---|---|---|\n| Multi-session reasoning | 4 / 5 |\n2 / 5 |\n| Knowledge update | 3 / 5 | 3 / 5 |\n| Single-session (user) | 3 / 5 | 3 / 5 |\n| Single-session (assistant) | 2 / 5 | 2 / 5 |\n| Single-session (preference) | 2 / 5 | 2 / 5 |\nTotal |\n14 / 25 |\n12 / 25 |\n\n**Storage:** Enki answers from**0.49× the stored facts** mem0 keeps on the same conversations (mean 138 vs 283).**Standout:** multi-session reasoning (4/5 vs 2/5).\n\nHonest framing.This is a small, hand-validated slice; the overall margin (14 vs 12) is modest and within what a 25-item sample can show. The robust, repeatable result iscomparable answer accuracy at roughly half the memory footprint, with a clear multi-session advantage. Further evaluation is ongoing.\n\nMeasured on a ~139-fact store, CPU-only (no GPU), 240 samples:\n\n| Percentile | Latency (ms) |\n|---|---|\n| mean | 7.6 |\n| p50 | 6.1 |\n| p95 | 11.9 |\n| p99 | 13.0 |\n\nFull methodology and per-question results are available on request.\n\n*Enki Labs (UK) · 2026*", "url": "https://wpnews.pro/news/enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well", "canonical_source": "https://github.com/stephen487/enki-benchmarks", "published_at": "2026-06-27 23:35:47+00:00", "updated_at": "2026-06-28 00:04:50.083843+00:00", "lang": "en", "topics": ["large-language-models", "ai-agents", "ai-research"], "entities": ["Enki Labs", "Enki", "mem0", "Claude Haiku", "LongMemEval-S"], "alternates": {"html": "https://wpnews.pro/news/enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well", "markdown": "https://wpnews.pro/news/enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well.md", "text": "https://wpnews.pro/news/enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well.txt", "jsonld": "https://wpnews.pro/news/enki-memory-for-ai-agents-that-keeps-half-as-much-and-answers-as-well.jsonld"}}