CoreCognition Single Image Mechanicalreasoning: leaderboard
CoreCognition Single Image Mechanicalreasoning — 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
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | qwen2.5-vl-72b-instruct_video | CoreCognition run: native-setting=qwen2.5-vl-72b-instruct_video | 85.66 | |
| 2 | LLaVA-Video-72B-Qwen2 | 77.42 | #1228 | |
| 3 | InternVL2.5-38B | 70.79 | #1413 | |
| 4 | VideoLLaMA2-72B | 69.71 | #1355 | |
| 5 | InternVL2.5-78B | 67.56 | #1044 | |
| 6 | Llama 3.2 90B Vision Instruct | 67.56 | #1769 | |
| 7 | llava-next-video-32B-qwen | 67.28 | ||
| 8 | LLaVA-Video-7B-Qwen2 | 63.68 | #1854 | |
| 9 | qwen2.5-vl-3b-instruct_video | CoreCognition run: native-setting=qwen2.5-vl-3b-instruct_video | 63.17 | |
| 10 | InternVL2.5-8B | 63.08 | #2444 | |
| 11 | qwen2.5-vl-7b-instruct_video | CoreCognition run: native-setting=qwen2.5-vl-7b-instruct_video | 62.99 | |
| 12 | LLaVA-CoT | 61.83 | #3271 | |
| 13 | InternVL2.5-26B | 61.56 | #2269 | |
| 14 | yi-vision-v2 | 60.39 | ||
| 15 | llava-video-7B-qwen2-video-only | 56.14 |
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-mechanicalreasoning · How It Works · Data refreshed daily, snapshot 2026-11-16.