CUE-Mem - Text Index Multimodal Use - Answer Refusal Explicit: leaderboard
CUE-Mem v1 RAG multimodal; source index/use modalities and explicit/implicit evidence kept separate. Oracle annotated clues bypass retrieval. Memory Average excludes refusal; other columns are four-choice task accuracy. Native interfaces differ; exact Qwen3OmniInstruct and MiniCPMo4.5 identities retained.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around February 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3 Omni 30B A3B Instruct | 45.6 | #401 |
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-index-multimodal-use-answer-refusal-explicit · How It Works · Data refreshed daily, snapshot 2026-11-05.