CoreCognition Image + Video Permanence: leaderboard

CoreCognition Image + Video 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. 120 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1LLaVA-NeXT-72B61.03#3185
2Gemini 1.5 Pro59.56#1061
3Gemma 3 4B (IT)58.66#4078
4Llama-3-VILA1.5-8B58.09#7424
5Gemini 1.5 Flash-8B58.01#2670
6O1 (2024-12-17)57.03#313
7llama-3-longvila-8b-1024framesCoreCognition run: native-setting=llama-3-longvila-8b-1024frames55.88
8llama-3-longvila-8b-256framesCoreCognition run: native-setting=llama-3-longvila-8b-256frames54.41
9MiniCPM-Llama3-V2.554.17#5149
10llama-3-longvila-8b-512framesCoreCognition run: native-setting=llama-3-longvila-8b-512frames54.17
11internvl-chat-v1-253.21
12InternLM-XComposer2.553.1#3981
13GPT-4o52.86#744
14Mini-InternVL-Chat-4B-V1.551.47#4881
15llama-3-longvila-8b-128framesCoreCognition run: native-setting=llama-3-longvila-8b-128frames50.98

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

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