Skip to content
MemJudge
LeaderboardCompareRunHow scoring worksGitHub

Seasons

A season is one fixed set of judging rules: the judge model, its settings and the judge prompt. Every result records its season, and the leaderboard ranks only results from the same season. When the rules change, a new season starts. Earlier runs stay listed below the ranking until they are re-judged, or run again when what the judge sees has changed. How seasons work.

Season 1CurrentProvisional

First season: Claude Sonnet 5.5 with judge prompt v3.

Started
28 Oct 2026
Judge
claude-sonnet-5-5, effort medium
Prompt
v3 sha256:ce6d9750331c…
Details
The change that started it ↗

MemJudge is a free, non-commercial, vendor-neutral benchmark for AI memory systems. The code is Apache-2.0 and scores are published under CC BY 4.0; dataset text in the results keeps its own licence (LoCoMo is CC BY-NC 4.0, non-commercial). It is built by Bit Robotics Ltd, which makes Recalld, so every run says whether its submitter is affiliated with the system it tests.

Every result here is signed. How signing works and how to check a result yourself.

Terms of Use · Privacy notice · Corrections policy · admin@memjudge.org

Live: site 4c60e4d · results ffc8d78