RVMS-Bench (RACLO) - Moment Localization (Key Moment, Temporal): leaderboard
Metric: Moment localization accuracy (%) on the 160 queries giving the key moment and temporal context: the frame the model predicts for the remembered moment is certified as matching the gold frame by a Gemini 2.5 Pro verifier; closed models see 32 to 192 frames by video length, open models 64; RVMS-Bench's real-world YouTube video search queries built from fuzzy memory cues (global impression G, key moment K, temporal context T, auditory memory A; 160 queries per cue combination, 20 topics, four duration bands), each model run as the backbone of the RACLO open-web search agent (two search queries, two candidate URLs each); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 38.8 | #105 |
| 2 | Gemini 3 Pro | 30 | #77 |
| 3 | Gemini 2.5 Pro | 18.8 | #145 |
| 4 | GPT-4o | 16.9 | #333 |
| 5 | GPT-5 Mini | 8.8 | #176 |
| 6 | Qwen 3 VL 235B A22B Instruct | 3.1 | #264 |
| 7 | Qwen 2.5 VL 72B Instruct | 0 | #364 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=rvms-bench-raclo-moment-localization-key-moment-temporal · How It Works · Data refreshed daily, snapshot 2026-10-11.