New Benchmark Pits Human Writers Against 24 LLMs Across 475 Prompts, Results Show ...

A new benchmark by Vulsar AI tested 24 large language models (LLMs) against amateur and professional human writers across 475 creative writing prompts. The results show that only the most advanced frontier models like GPT 6 Astra slightly outperform amateur humans, while professional writers still outperform all AI tested.

This Wire brief sits within Fusion42's coverage of Generative AI and AI Frontier Models.