Benchmark reproducibility & verification status

July 15, 2026 · View on GitHub

FluctlightDB publishes benchmark numbers in the paper, README, and frozen JSON under benchmarks/results/. This document separates what you can reproduce yourself from what has been independently verified.

Why independent reproduction benefits us

BenefitWhat it unlocks
CredibilityMoves headline metrics from “maintainer blog post” to “anyone can verify” — rare in the agent-memory space
AdoptionEnterprise / research integrators can pin versions and run make reproduce-locomo in CI before trusting recall claims
DifferentiationHonest REPRODUCIBILITY.md + open harness vs competitors’ self-reported leaderboard percentages
Co-maintainersExternal reproducers are the best pipeline for triage contributors and named reviewers
Head-to-head postsMem0/Zep comparisons stay blocked until the first external LoCoMo match — then a protocol-identical post is credible, not more noise
Paper & press“Independently reproduced by X” is citable; a self-reported number alone is not

We welcome independent reproductions and list public credit in this doc (no cash bounty — maintainer-funded rewards are not offered).

Will you get the same %?

Honest LoCoMo headline: 96.8% @150 (MiniLM) / 97.0% (mpnet) raw evidence recall, no expansion — tight-k @5=72.6%/75.1%, @10=80.0%/82.6%, @20=85.6%/87.2%, @50=91.8%/92.4%. Native Rust CHORUS first-principles invented stack; bench benchmarks/locomo_engine_maxsim.py, frozen benchmarks/results/locomo-invented-stack-engine-2026-07-13.json (MiniLM) / locomo-mpnet-engine-2026-07-15.json (mpnet). The legacy make reproduce-locomo protocol below reproduces the deprecated ±3-expansion cert (the old 99.0%), kept for historical parity only. Evidence recall ≠ QA accuracy (E2E ≈85% @k=15).

Legacy expansion cert (deprecated) — CHORUS @ k=150: yes — when you follow the pinned protocol.

make reproduce-locomo exits 0 only if your output JSON matches the frozen cert exactly on:

  • evidence_hits (string, e.g. "1970/1982")
  • mean_evidence_recall (float, e.g. 0.990201277587756)

That is the legacy 99.0% (1970/1982) ±3-expansion cert — deprecated, not the headline (the honest number is 96.8% @150 no-expansion, above).

RequirementWhy it matters
Git tag matching the cert (e.g. v0.5.6)Recall logic lives in fluctlightdb-native
fluctlightdb[native]==<tag version>Script pins from crates/fluctlight-py/pyproject.toml
benchmarks/requirements-reproduce.txtPins Chroma / eval deps (embeddings)
--mode chorus --top-k 150Default in reproduce-locomo.sh
FLUCTLIGHT_FABRIC=1Paper-profile Recall Fabric on (set by reproduce-locomo.sh; also via configure_ir_env() in locomo_eval.py)
Frozen cert (honest raw, no expansion)benchmarks/results/locomo-lateinteraction-2026-07-13.json (self-contained); engine: locomo-invented-stack-engine-2026-07-13.json (MiniLM) / locomo-mpnet-engine-2026-07-15.json (mpnet)
Official locomo10.jsonAuto-downloaded from snap-research/locomo

What can still differ (without changing the pass/fail check): wall_s, embed cache hit counts, absolute paths in JSON metadata.

What can cause a mismatch:

CauseTypical symptom
Unpinned / newer chromadb (ONNX MiniLM drift)evidence_hits off by a few spans
Wrong native version or building from arbitrary mainRecall ranking changes
Wrong lane (agent vs chorus) or top_kVery different %
Modified dataset fileAny divergence

LongMemEval / BEIR: More moving parts (mpnet server, Chroma versions, rerank flags). Exact float match is harder; report your numbers via the reproduction issue template even if not bit-identical.

We have not yet had an external party confirm a match — if you reproduce (or disprove) a frozen cert, use the issue template below.


Leaderboards (industry context)

There is no canonical live leaderboard for LoCoMo or LongMemEval maintained by a neutral third party. Headline numbers from Mem0, Zep, ByteRover, Hindsight, Memobase, etc. are overwhelmingly vendor blog posts or self-published tables, not audited rankings.

FluctlightDB's differentiation is honesty + open harnesses:

  • Frozen JSON + make reproduce-locomo (not just a percentage in a README)
  • This doc states self-reported until independently verified
  • No Mem0/Zep head-to-head post until an external reproduction lands (would add noise before that)

Independent reproduction (public credit)

BenchmarkFirst external reproductionRecognition
LoCoMo CHORUS @150Not yetPermanent credit in this doc + README
LongMemEval-S session@8Not yetSame
BEIR SciFact nDCG@10Not yetSame
FAMB macroNot yetSame

LongMemEval E2E is excluded (locked maintainer run; OpenAI cost).

How to report: open a Benchmark reproduction issue, include environment + output JSON. Maintainer verifies and updates the table below.

Confirmed external reproductions

DateWhoBenchmarkResultMatch
none yet

Verification status (honest summary)

StatusMeaning
Open harnessScript/notebook + frozen JSON exist in this repo; anyone can re-run
Self-reportedNumbers were produced by the maintainer; no third party has published a matching run
LockedRe-run is intentionally frozen (cost, API keys, or paper freeze)
BenchmarkFrozen artifactOpen harnessMaintainer self-reportedIndependent third-party reproduction
LoCoMo evidence recall @150 (honest raw, no expansion)locomo-invented-stack-engine-2026-07-13.json (MiniLM), locomo-mpnet-engine-2026-07-15.json (mpnet)benchmarks/locomo_engine_maxsim.pyYes — 96.8% @150 (MiniLM) / 97.0% (mpnet), @5=72.6%/75.1% tight-kNone published
LongMemEval-S session@8paper-2026-07-09.jsonColab + local scriptsYes — 97.6%None published
LongMemEval E2E QAe2e-cert-paper-v2-2026-07-07.jsonbenchmarks/e2e_certify.sh (OpenAI)Yes — 97.4%None — run locked
BEIR SciFact nDCG@10 / R@10paper-2026-07-09.jsonbenchmarks/beir_bench.pyYes — 0.646 / 0.792 (Fabric on)None published
FAMB macropaper-2026-07-09.jsonbenchmarks/famb_bench.pyYes — 100%None published

Bottom line: Harnesses are open and numbers are frozen, but all headline metrics are maintainer-reported until an external group publishes a reproduction (issue, blog, paper, or fork).

Reproduce LoCoMo (fastest independent check)

git clone https://github.com/voxmastery/FluctlightDB.git && cd FluctlightDB
git checkout v0.5.6   # or tag matching the frozen cert you compare against
make reproduce-locomo

This downloads LoCoMo data, runs the honest self-contained recipe (benchmarks/locomo_lateinteraction.py — token-population MaxSim ⊕ BM25, no neighbor expansion), and compares raw recall@150 against locomo-lateinteraction-2026-07-13.json. The native-engine invented-stack headline (96.8% @150 MiniLM / 97.0% mpnet, no expansion) reproduces via benchmarks/locomo_engine_maxsim.py after a source build, comparing against locomo-invented-stack-engine-2026-07-13.json / locomo-mpnet-engine-2026-07-15.json. The old ±3-expansion 99.0% is deprecated (a trivial BM25 baseline also scored ~99% under it).

Options:

REPRODUCE_FROM_SOURCE=1 make reproduce-locomo   # build native wheel from source
OUT=benchmarks/results/my-run.json bash scripts/reproduce-locomo.sh
pip install -r benchmarks/requirements-reproduce.txt   # deps only

Other benchmarks

BenchmarkHow to reproduceNotes
LongMemEval-Sbenchmarks/longmemeval_colab_v2.ipynb or local script in benchmarks/README.mdRequires embedding server (mpnet) for local path
LongMemEval E2EE2E_PROFILE=paper benchmarks/e2e_certify.shOpenAI API cost; maintainer locked paper run — do not overwrite e2e-cert-paper-v2-2026-07-07.json
BEIR SciFactbenchmarks/beir_bench.pyChroma + pytrec_eval deps
Full paper bundlebenchmarks/results/paper-2026-07-09.jsonAggregates frozen runs

Full protocol: BENCHMARKS.md · benchmarks/README.md

Metric caveats (read before comparing to Mem0/Zep)

  • LoCoMo evidence recallLLM-judge E2E QA used by some memory SDK leaderboards. Different task, different numbers.
  • Embeddings: paper uses all-MiniLM-L6-v2 (ONNX CPU) unless noted; LongMemEval v4 path uses mpnet.
  • CHORUS lane uses bulk imprint (connect_chorus()); agent lane uses connect_agent().

Report your reproduction

Use the Benchmark reproduction issue template (label: reproduction).

Include: commit hash, OS, Python version, fluctlightdb / fluctlightdb-native versions, output JSON path, diff vs frozen cert.

Partial reproductions (e.g. within ±0.5% due to embedding drift) are still valuable — report them.

What we do not claim

  • Peer review of benchmark methodology (paper is preprint)
  • Independent audit of correctness or security
  • Leaderboard registration on a third-party site — see LEADERBOARD.md (no neutral agent-memory registry exists)