CoreCognition Single Image Permanence: leaderboard

CoreCognition Single Image Permanence — CoreCognition native component accuracy and reported mean. Image+Video, image-only and single-image populations are separate; no pooling or replacement across cohorts. Metric: Accuracy (%). Source: https://grow-ai-like-a-child.github.io/CoreCognition/leaderboard

Metric: Accuracy (%). Source: grow-ai-like-a-child.github.io. Saturation forecast: Estimated already saturated. 33 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1videochat2-stage3-mistral-7B48.69
2videollama2-7B-16f45.26
3chat-univi-13B45.26
4chat-univi44.61
5LLaVA-CoT43.95#3271
6LLaVA-Video-72B-Qwen242.89#1228
7InternVL2.5-78B41.09#1044
8InternVL2.5-8B40.85#2444
9VideoLLaMA2-7B40.85#2213
10video-chat-13B40.85
11qwen2.5-vl-7b-instruct_videoCoreCognition run: native-setting=qwen2.5-vl-7b-instruct_video40.2
12qwen2.5-vl-72b-instruct_videoCoreCognition run: native-setting=qwen2.5-vl-72b-instruct_video38.81
13VideoLLaMA2-72B38.24#1355
14llava-video-7B-qwen2-video-only37.99
15InternVL2.5-38B37.91#1413

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=corecognition-single-image-permanence · How It Works · Data refreshed daily, snapshot 2026-11-16.