OpenAI GPT-o3: Scores 87.5% on AGI Benchmark, Solves PhD-Level Problems
OpenAI o3 scores 87.5% on ARC-AGI (near human level), 96.7% on AIME math, and 71.7% on SWE-bench coding benchmark.
Latest stories, in-depth analysis & breaking updates
OpenAI o3 scores 87.5% on ARC-AGI (near human level), 96.7% on AIME math, and 71.7% on SWE-bench coding benchmark.