by nopinsight on 4/22/25, 3:28 PM with 1 comments
by nopinsight on 4/22/25, 3:31 PM
OpenAI’s o3 now outperforms 94% of expert virologists." -- thread by a co-author, https://x.com/DanHendrycks/status/1914696657813561799
Paper: Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark