Grok 4.5 ranks #1# overall on the HighWalk benchmark, outperforming Claude Opus 5, Fable 5 and GPT-5.6 Sol It delivered the highest quality scores while taking just 1m 52s and costing only $0.25....that’s significantly less For comparison: • Claude Opus 5: 4m 34s and $1.87 • Claude Fable 5: 4m 49s and $2.87 • GPT-5.6 Sol: 16m 49s and $10.14 Grok didn’t win by throwing more time, tokens and money at the task. It won by delivering the best overall balance of intelligence, speed and cost That is what makes the most useful frontier AI

GPT-5.4 helped drive a medicinal chemistry project from literature review to a validated experimental result. Paired with Maria AI and specialized lab, the model proposed an unexpected way to improve a widely used reaction in drug discovery.