HUSH-Bench (MemU): leaderboard

Metric: Unsolicited history integration score (0-100): a GPT-5.4 judge scores use of prior sensitive history the current benign prompt gives no reason to use; 2,400 probes per cell; temperature 0, 800-token budget; lower is better; memory entries retrieved by MemU (up to five). Source: arxiv.org. Saturation forecast: Around December 2027. 4 models tracked.

Top models

#ModelScore
1GPT-5.4 Mini26.6
2Claude Sonnet 4.670.6
3DeepSeek V4 Flash79.5
4Qwen 3.5 9B81.4

Interactive version: theaggregate.ai/benchmark?slug=hush-bench-memu · How It Works · Data refreshed daily, snapshot 2026-09-29.