OpenAI: GPT-5 demonstrates ability to assist research in a real-world laboratory environment

According to the AXIOS website, OpenAI has revealed for the first time that GPT-5 has demonstrated the ability to assist scientific research in a real laboratory environment, indicating that artificial intelligence is expected to play a more central role in scientific experiments. While AI is advancing rapidly in areas such as mathematics and physics, major breakthroughs in biology have been relatively slow, as biology is highly dependent on real-world laboratory work, not just computational simulations. OpenAI has teamed up with biosafety start-up Red Queen Bio to build a testing framework to evaluate how AI models perform in the lab. The lab involves "wet" operations such as liquids, chemicals and biological samples to distinguish it from "dry" labs that focus on data analytics. In the experiment, GPT-5 proposed improvements to the molecular cloning experimental process. After the human scientists executed it, the results were fed back to GPT-5, and GPT-5 iteratively optimized the scheme based on the feedback. The experimental results showed that GPT-5 increased the efficiency of a standard molecular cloning process by 79 times.