VGenST-Bench - Interacted Object Identification: leaderboard

VGenST-Bench (2026) tests spatio-temporal reasoning of multimodal LLMs on short videos synthesized from validated scene graphs with text-to-image and image-to-video models. Interacted Object Identification task: identify which object an agent picked up and moved (vista scale, egocentric, dynamic). Accuracy in percent on multiple-choice questions under circular evaluation (right under every rotation of the options), 8 sampled frames; higher is better.

Metric: Interacted Object Identification (vista scale, egocentric, dynamic) accuracy (%) on the base multiple-choice questions of VGenST-Bench under circular evaluation (correct only if right under every cyclic permutation of the options), macro-averaged over the task's QA types, 8 uniformly sampled frames from about 100 synthesized videos per task (generated with text-to-image and image-to-video models from validated scene graphs); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.494#135
2Gemini 3 Flash93.9#175
3Gemini 3.1 Flash Lite93.9#371
4Qwen 3.5 9B92.1#828
5Qwen 3.5 4B91.9#1241
6Qwen 3.5 27B91.4#375
7Kimi K2.687.6#192
8Gemma 4 31B (IT)82.2#350
9GLM-4.6V-Flash80.5#1040
10GPT-5.4 Mini80.4#406
11Gemma 4 26B A4B (IT)77.2#556
12InternVL3.5-30B-A3B74.9#1267
13GPT-5.4 Nano62.9#651
14InternVL3.5-8B60.2#1604
15InternVL3.5-4B55#1930

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=vgenst-bench-interacted-object-identification · How It Works · Data refreshed daily, snapshot 2026-11-21.