OpenAI says planned GPT-6.1 is too insecure to release (arstechnica.com)

by rss-bot · 1 month ago · 0 comments
Similar performance, security tradeoffs also seen in current public models.