Will Depue Questions GPT-6 Launch Without Evals
OpenAI engineer Will Depue flags missing benchmark results in early GPT-6 coverage.
TLDR
Will Depue, a Research Engineer who previously worked at OpenAI on Sora and o-series models, posted on X that he does not understand the GPT-6 launch. He noted that a media embargo had lifted, CNET had begun coverage, and OpenAI staff were tweeting, yet no evaluations appeared. The post treats the absence of benchmark data as the immediate point of confusion among visible replies.
Combined views
1.7K
1 Source, first seen 27d ago