Multi-Model Ensemble Hits 72% on ARC-AGI-2 Benchmark

A parallel ensemble of GPT-5.2, Gemini 3, and Claude Opus 4.5 has scored 72% on ARC-AGI-2 — the highest reported result yet, and a data point for multi-model architectures as a path to reliable generalization.

A new result on the ARC-AGI-2 benchmark — widely considered one of the hardest tests of general reasoning — has reached 72% by running GPT-5.2, Gemini 3, and Claude Opus 4.5 in parallel and aggregating their outputs. As @testingcatalog reported, no single model achieves this score alone; it's the ensemble that pushes the frontier.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Vestibulum ante ipsum primis in faucibus orci luctus et ultrices posuere cubilia curae.

Unlock the full briefing

Get every story in today's briefing, the full archive, and the daily AI intelligence brief.

All stories today

Full archive

Daily brief