MIRIX-2.0 multi-agent memory system tops LoCoMo at 91.4%, beats best baseline by 14.3 points
_akhaliq · x · 2026-11-17
MIRIX-2.0 is a practical multi-agent memory system for LLM agents that combines factual and procedural memory.
- Scores 91.4% on LoCoMo
- Scores 72.6% on LongMemEval-S
- Leads the strongest baseline by 14.3 points
The results were amplified by the HuggingPapers account, making it one of the leading open-source works on long-term agent memory.
More from coding & agent
- Google's Nov 10 Core Update: early data shows YMYL site down ~60%, affiliate template pages hit — gaganghotra_ · 2026-11-17
- "No AI Slop Manifesto" passes 660 signatures from OpenAI, SpaceX builders — Madisonkanna · 2026-11-17
- Nasiko launches open runtime for agents, cutting input tokens 42% via tool-output compression — socialwithaayan · 2026-11-17
- Same model, 15x more broken tool calls: dottxt's provider tool-call index — remilouf · 2026-11-17
- Same prompt showdown: Claude vs ChatGPT vs Figma Make UI generation compared — Tegadesigns · 2026-11-17
- LangChain webinar: Harrison Chase on the future of agents, Friday Nov 20 — LangChain · 2026-11-17