benchmark · first seen 4 Oct 2026
LongMemEval
A benchmark evaluating long-term conversational memory in chat assistants across multi-session and temporal questions.
Mentions per week
PapersCodeDiscussion
Earliest sources
- MemStrata: 95% and 90.91% Source-Aware Accuracy on LongMemEval-500 and LoCoMo-1540 with a Local Qwen 3.8 27B Q4_K_M ReaderPapers · 4 Oct 2026 · Neeraj Yadav
- MINDSET: Energy-based Memory Evolution for Long Conversational AgentsPapers · 6 Oct 2026 · Sujato Dutta, Sreekruthy Tummala, Shashank Vanga, Ayushmi Pavani, BHVSP Subrahma…
Latest sources
- MINDSET: Energy-based Memory Evolution for Long Conversational AgentsPapers · 6 Oct 2026 · Sujato Dutta, Sreekruthy Tummala, Shashank Vanga, Ayushmi Pavani, BHVSP Subrahma…
- MemStrata: 95% and 90.91% Source-Aware Accuracy on LongMemEval-500 and LoCoMo-1540 with a Local Qwen 3.8 27B Q4_K_M ReaderPapers · 4 Oct 2026 · Neeraj Yadav