Sublime
An inspiration engine for ideas
We’re releasing PaperBench, a benchmark evaluating the ability of AI agents to replicate state-of-the-art AI research, as part of our Preparedness Framework.
Agents must replicate top ICML 2024 papers, including understanding the paper, writing code, and executing experiments. https://t.co/CvYcDdk0nI
Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date.
For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation. https://t.co/rDaqV0x0wE
OpenAIx.com
🆕It's @openai o3-pro launch day!
our high taste guest tester @benhylak has been previewing for the past week and found an interesting pattern: (link in reply)
o3-pro doesn't noticeably outperform in normal situations, but it's just really, really, REALLY good at consuming ALL your context... See more
OpenAI dropped their GPT-5 roadmap. They're going all in on making one super-model that does everything. GPT-4.5 drops in a few weeks (last basic model) then the real deal GPT-5 comes in a few months, basically merging o3 tech into it. Pretty wild - it'll figure out on its own when to use stuff like Voice and Canvas. The best part? Everyone gets... See more
OpenAI
openai.com
