GPT-5.4 Passed the Human Benchmark for Desktop Tasks — What It Means

May 2 · 9 min read · On April 4, 2026, OpenAI published benchmark results showing GPT-5.4 scored 75.0% on OSWorld-Verified — surpassing the human baseline of 72.4%. That 2.6 percentage point margin may look modest, but the context makes it significant: GPT-5.2, the previ...