Back to benchmarks

Research

LoCoMo self-run
notes

What we measured, how we measured it, and what we are not claiming. Numbers verified against the raw result JSON from 2026-07-10.

Headline

v0

31.7%

488/1540

v0.2

77.9%

1200/1540

Same dataset and question set both runs. Jump is methodology parity with the open eval stack plus extractor fixes — not a different dataset.

Method

Dataset
LoCoMo-10 (non-adversarial)
Questions
1540 (adversarial category excluded)
Harness
mem0 open eval harness (memory-benchmarks-style)
Extract
gpt-4o-mini
Answer
gpt-4o-mini
Judge
gpt-4o-mini
Embeddings
text-embedding-3-small
Recall budget
6,500 tokens · k=150
Store
InMemoryBeliefStore (accuracy focus; latency not claimed)

By category

Grey = v0 · Red = v0.2. Fractions are correct / total in that category.

Multi-hopv0 39/282 · v0.2 235/282
v0
13.8%
v0.2
83.3%
Temporalv0 127/321 · v0.2 239/321
v0
39.6%
v0.2
74.5%
Open-domainv0 12/96 · v0.2 71/96
v0
12.5%
v0.2
74%
Single-hopv0 310/841 · v0.2 655/841
v0
36.9%
v0.2
77.9%

Product suite (our wedge)

On a separate, small harness aimed at staleness and conflict (location update, preference flip, job change, multi-fact, scope isolation): mem01 5/5. Same harness, mem0 OSS scored 2/5.

Internal staleness/conflict product suite (location flip, preference flip, job change, multi-fact, scope isolation). This is not LoCoMo; it is the failure mode we optimize for.

What we are not claiming

  • mem0’s published ~92.5% LoCoMo figure uses a stronger gpt-4o-class answer/judge stack. Our self-run uses gpt-4o-mini throughout for cost and reproducibility.
  • Single full run; LLM judge noise is roughly ±2 points on 1,540 questions.
  • In-memory store for this accuracy run — we do not claim production latency numbers from it.
  • Per-query token/latency telemetry was not recorded for this run (backlog).

Bottom line

Under a fixed, disclosed stack (gpt-4o-mini everywhere), mem01 moved from 31.7% → 77.9% on LoCoMo-10. Gains show up across multi-hop, temporal, open-domain, and single-hop. Headline vendor numbers with stronger models are a different comparison — we keep our model choice public.

Back to site