← All systems

Zep (Graphiti)

OrganizationZep AI (YC W24)
Founded2023
FundingYC $500K verified; no later round found
CodeApache-2.0 · 29.2k★

Bi-temporal knowledge graph (Graphiti) with hybrid semantic + BM25 + graph retrieval.

Every number on file

claimedbenchmarkconfigrunnerstatus
71.2 / 63.8LongMemEval-Sgpt-4o / 4o-mini readerself (paper)SELF-REPORTED
75.14 ± 0.17LoCoMocorrected downward from ~84 after disputeself (blog)DISPUTED
94.7 / 90.2LoCoMo / LMEgpt-5.4 reader + gpt-5.4 CoT judge · undatedself (research page)SELF-REPORTED
58.44 ± 0.20LoCoMo10 runs, Cat-5 excludedMem0 re-runDISPUTED
DNF (>2 days)LongMemEvalunified harness, Qwen2.5-7Bsurvey (arXiv:2604.01707)THIRD-PARTY

The only vendor that publicly corrected its own number (75.14). The 2025 dispute with Mem0 is the canonical case of config-dependent scoring.

Rows are not comparable — configs differ. A row graduates to a verdict only through a registered re-run. The protocol →