TOC-Bench - Reappearance or Disappearance: leaderboard
The reappearance or disappearance questions of TOC-Bench (2026), video questions grounded in per-frame object tracks and kept only when not answerable from text alone, single frames or shuffled frames: whether an object that vanished was fully blocked from view or left the frame. Accuracy in percent; higher is better.
Metric: Reappearance or Disappearance accuracy (%) on TOC-Bench (2,323 human-verified questions over 1,951 videos from Charades, Perception Test, OVIS and MOSE, grounded in per-frame object tracks and kept only when not answerable from text alone, single frames or shuffled frames), mostly 32 uniformly sampled frames, deterministic scoring of multiple-choice, statement-pair, numeric and ordering formats; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 21 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 78.4 | #56 |
| 2 | Grok 4.3 | 70.9 | #351 |
| 3 | Kimi K2.6 | 69.4 | #197 |
| 4 | GLM-5V Turbo | 67.9 | #365 |
| 5 | Gemini 3.1 Pro (Preview) | 67.2 | #102 |
| 6 | Video-LLaVA-7B | 59.7 | #3904 |
| 7 | Gemini 3.1 Flash Lite (Preview) | 58.2 | #419 |
| 8 | MiMo-V2-Omni | 57.5 | #465 |
| 9 | Seed 2.0 Lite | 49.3 | #306 |
| 10 | Qwen 2.5 VL 72B | 49.3 | #782 |
| 11 | GPT-5.4 Mini | 47.8 | #409 |
| 12 | Video-ChatGPT | 47.8 | #6949 |
| 13 | VideoChat-R1.5-7B | 47 | #1260 |
| 14 | LLaVA-Video-72B | 47 | #1211 |
| 15 | MiniCPM-V-2.6 | 44 | #3311 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=toc-bench-reappearance-or-disappearance · How It Works · Data refreshed daily, snapshot 2026-11-22.