RuntimeWire ranks nine AI models on news research, summaries and writing

GPT-6.1 Sol Pro leads the equal-weight index, which combines three task suites with samples ranging from 50 to 100 tasks.

By · Published · Updated

A researcher prepares to run a software benchmark on an unbranded laptop at a workbench.

GPT-6.1 Sol Pro leads the equal-weight index, which combines three task suites with samples ranging from 50 to 100 tasks.