CUE-Mem - Text Reflexion - Entity Recall Explicit: leaderboard
CUE-Mem v1 textualized memory; four-choice accuracy; source method and explicit/implicit evidence separate; Memory Average excludes refusal. Oracle uses synthesis source prompts, not retrieved evidence. No Context is a diagnostic control. Full Memory4096words; retrieval methods retain their own documented storage/update policies. Qwen3.6-35B-A3B and GPT5.4mini source identities; no derived EI-gap.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around February 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 35B A3B | 80.5 | #238 |
| 2 | GPT-5.4 Mini | 75.3 | #228 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=cue-mem-text-reflexion-entity-recall-explicit · How It Works · Data refreshed daily, snapshot 2026-11-05.