OpenAI introduces GPT-5.4 with more knowledge work capability
SMRTR summary
OpenAI launched GPT-5.4 amid growing competition from Anthropic, which gained users after OpenAI's controversial Pentagon deal. The new model features enhanced visual analysis capabilities for images up to 10.24 million pixels and produces 18 percent fewer factual errors than previous versions.
SMRTR provides this summary for quick context. The original article belongs to Ars Technica.
Read the original article