VGenST-Bench - Behavioral Trigger 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. Behavioral Trigger Identification task: identify the moving hazard that made an agent stop, not the static distractor (environmental scale, exocentric, 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: Behavioral Trigger Identification (environmental scale, exocentric, 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
1Gemini 3 Flash96.9#175
2InternVL3.5-30B-A3B90.4#1267
3GPT-5.490.1#135
4Gemini 3.1 Flash Lite88#371
5Gemma 4 31B (IT)84.6#350
6Kimi K2.680.9#192
7GPT-5.4 Mini80.2#406
8Gemma 4 26B A4B (IT)79.7#556
9GPT-5.4 Nano78.8#651
10GLM-4.6V-Flash76.9#1040
11InternVL3.5-8B76.4#1604
12InternVL3.5-4B70.2#1930
13Qwen 3.5 27B64.8#375
14Qwen 3.5 9B63.9#828
15Qwen 3.5 4B58.9#1241

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-behavioral-trigger-identification · How It Works · Data refreshed daily, snapshot 2026-11-21.