OpenAI just launched o1, the first in a brand-new series of “reasoning” models that are trained to handle more complex questions—faster than humans! 😲 Along with this, they’ve released o1-mini, a smaller, more affordable version. (Yes, AI insiders, this is the much-hyped “Strawberry model” 🍓.)
So, what’s the buzz? 🎉 o1 is a game-changer for writing code and solving multistep problems with greater accuracy than ever before. It’s a step toward OpenAI’s ultimate goal of human-like intelligence. However, it’s still in preview, meaning it’s early in development and slightly pricier (and slower) than GPT-4o.
Developers: ⚙️ The API access for o1 comes with premium pricing at $15 per 1 million input tokens and $60 per 1 million output tokens (vs GPT-4o’s $5 and $15). But with fewer hallucinations and sharper reasoning, this model is definitely a power move! 💪
Here’s my summary of the key wow points:
