CoreCognition Image only Perceptualconstancy: leaderboard

CoreCognition Image only Perceptualconstancy — 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: Rough model projection: around 2026. 77 models tracked. Updated 16 Nov 2026.

Top models

#ModelScoreOverall rank
1GLM-4V-Plus80.42#1400
2Cambrian-8B77.62#6834
3step-1o-turbo-vision76.69
4Cambrian-34B76.46#5295
5IDEFICS2-8B74.83#5096
6Step-1.5V-mini73.19#2720
7360VL-70B72.03#5987
8InternLM-XComposer2-4KHD69.7#6477
9InternLM-XComposer2-1.8B69.23#8431
10PaliGemma-3B-mix-44868.65#6985
11Phi-3-Vision68.3#5517
12points-qwen-2.5-7B-chat68.3
13NVLM-D-72B68.18#2181
14wemm-chat-2k-cn67.37
15Monkey-Chat67.13#7874

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