# xiaowu0162/LongMemEval

Benchmarking Chat Assistants on Long-Term Interactive Memory (ICLR 2025)

Repository: https://github.com/xiaowu0162/LongMemEval
Canonical: https://ross.abutalabs.com/products/longmemeval
Language: Python
License: MIT
License Family: permissive
Last push: 2026-05-11T22:49:24+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 81, release rhythm 35, longevity 49
- inputs: {"age_days": 692, "days_push": 114, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1035, forks 81 (observed 2026-08-28T04:03:19.063223+00:00)

## What it is
LongMemEval is a benchmark of 500 high-quality questions for evaluating the long-term memory abilities of chat assistants across five skills including information extraction, multi-session reasoning, knowledge updates, temporal reasoning, and abstention. The repository provides the dataset, evaluation code, and tooling for running chat systems against timestamped multi-session chat histories.

## Use cases
- evaluate long-term memory of chat assistants
- benchmark llm memory across multiple sessions
- test temporal reasoning in conversational ai
- measure knowledge update handling in chatbots
- compare rag memory systems on long chat histories
- evaluate abstention when information is missing

## When to choose
- you need a standardized benchmark for long-term conversational memory
- you are building or comparing memory-augmented chat assistants
- you want needle-in-a-haystack style evaluation over multi-session chat histories

## When to avoid
- you need a general-purpose LLM benchmark unrelated to memory
- you want a training dataset rather than an evaluation benchmark
- you need agentic long-term memory evaluation, for which LongMemEval-V2 is more appropriate

## Facets
- artifact type: dataset
- maturity: active
- function: benchmarking, rag, llm-inference, chatbot
- domain: large-language-models, artificial-intelligence, chatbots, testing
- platform: python, cli
- tags: long-term-memory, chat-assistants, iclr-2025, needle-in-a-haystack, memory-benchmark, multi-session-dialogue, evaluation, retrieval-augmented-generation, natural-language-processing

## Member repositories
- xiaowu0162/LongMemEval (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:19.063223+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:05:10.718689+00:00, confidence not recorded.
  - readme: https://github.com/xiaowu0162/LongMemEval (fetched 2026-08-28T04:03:19.063223+00:00, sha 90615d215ad9)
- Data as of 2026-08-30T08:39:29.467469+00:00.