# AutoTrustAI/PaperGuru-Benchmark

Lifecycle-Aware Memory for long-horizon LLM agents — 66.05% on PaperBench, 94.66% on SurveyBench, 10 peer-reviewed acceptances at FSE/ICML/TOSEM/AEI/ICoGB

Repository: https://github.com/AutoTrustAI/PaperGuru-Benchmark
Canonical: https://ross.abutalabs.com/products/paperguru-benchmark
Language: TeX
License: NOASSERTION
License Family: other
Last push: 2026-06-08T21:14:07+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 86, release rhythm 35, longevity 8
- inputs: {"age_days": 117, "days_push": 86, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1324, forks 196 (observed 2026-08-28T04:04:22.346409+00:00)

## What it is
PaperGuru is a benchmark and research repository for Lifecycle-Aware Memory (LAM), a long-term memory primitive for long-horizon LLM agents, reporting state-of-the-art results on PaperBench (66.05%) and SurveyBench (94.66%). The repository contains benchmark materials, papers, and figure-reproduction assets written largely in TeX.

## Use cases
- evaluate long-term memory systems for LLM agents
- benchmark agents on long-horizon research tasks like paper reproduction
- compare memory architectures against PaperBench and SurveyBench baselines
- research lifecycle-aware memory for multi-step agent workflows
- reproduce figures and results from the PaperGuru papers
- study agent performance on academic survey generation

## When to choose
- you need rigorous benchmarks for long-horizon LLM agent memory
- you are researching memory primitives for multi-step agents
- you want to compare your agent against published PaperBench/SurveyBench results

## When to avoid
- you need a production-ready memory library or SDK to drop into your app
- you want a simple plug-and-play vector store or RAG tool
- you need a permissively licensed codebase (license is unclear/NOASSERTION)

## Facets
- artifact type: dataset
- maturity: active
- function: benchmarking, rag, llm-inference, agent-framework, data-science
- domain: artificial-intelligence, large-language-models, education
- platform: python, cross-platform
- tags: long-horizon-agents, lifecycle-aware-memory, paperbench, surveybench, academic-research, evaluation-benchmark, memory-systems, ai-agents, retrieval-augmented-generation, research

## Member repositories
- AutoTrustAI/PaperGuru-Benchmark (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:22.346409+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 2026-08-30T04:46:40.156441+00:00, confidence not recorded.
  - readme: https://github.com/AutoTrustAI/PaperGuru-Benchmark (fetched 2026-08-28T04:04:22.346409+00:00, sha d2432cc7a278)
- Data as of 2026-08-30T08:39:29.467469+00:00.