🧙🏼 OpenAI launches o1 🍓

Also: o1 demos + a prompting guide

in partnership with

Subscribe‎‎‏‏‎|Sponsor‏‎|1,000 best GPTs |Our top pick tools

Howdy, wizards.

To welcome the weekend, OpenAI finally dropped their new model, named o1 (codenamed Strawberry).

Here’s what you need to know:

Dario’s Picks – OpenAI o1 special edition

  1. OpenAI launches the o1 series of models, including o1-preview and o1-mini. They’re designed to spend more time thinking before they respond, which enables them to reason through complex tasks and problems.

     

     

    The models are already available in ChatGPT for Plus and Team users. You’ll notice how the model takes its time to reason through chain-of-though before responding.

     

    Sam Altman said this marks the beginning of a new paradigm, where AI can do general-purpose complex reasoning. And get this – the o1 model rank in the 89th percentile for competitive programming and exceeds PhD-level accuracy on benchmarks in physics, biology and chemistry.

     

    There’s a caveat though: there’s weekly limits of 30 messages for o1-preview and 50 for o1-mini. This will probably increase over time. Browsing, file and image uploads aren’t currently available – but will also be introduced later.

     

    In addition to the main o1 model, OpenAI also released the somewhat faster version named o1-mini for when extensive knowledge isn’t required. The mini version is also cheaper to use via API, for those building AI into their products.

     

    ‎ Why it matters‎ ‎ I personally can’t wait to start using these new models in my day to day work. My first impression is that it’s super cool to see the reasoning steps it uses, and I don’t think the slowness is too much of a problem if it can give me better answers. However, I see myself still relying on GPT-4o for two reasons: 1) the low weekly message cap on o1, 2) I typically drop dozens of screenshots to ChatGPT every day, and I can’t do that with the new model until it gets image and file uploads.

Read on for DEMOS and OpenAI’s tips for how to best prompt the new model.

This issue is brought to you by

“I’ll just do it myself.” Sound familiar?

If that’s your response as you stare at your endless to-do list, remember this: If you don’t have an assistant, you are the assistant. You need to delegate those tasks to a U.S.-based BELAY Virtual Assistant – and here are 25 things they could start to own for you on day one. In as little as one week, you can be intentionally matched with a BELAY Virtual Assistant, so you can get back to what only you can do: Growing your organization.

  1. Preview videos of o1 doing its thing. OpenAI has created some short, illustrative videos on what makes the new model different. I’ve highlighted the coolest ones below.

via OpenAI on X.com

From our partners

200+ hours of research on AI tools, prompting techniques & hacks packed in a solid 3 hour masterclass.

  1. Prompting GPT-4o. Especially relevant to those planning to use the new model via API, OpenAI shared some docs on how to best approach this, especially since the thinking (“reasoning tokens”) eats up a bit of the context window, which could make the model confused on long texts.

  1. *Superfilter manages your inbox like an expert assistant. Spend an extra hour a day enjoying your life instead of fighting for inbox zero.

     

    Try Superfilter today.

     

    *sponsored

Was this email forwarded to you? Sign up here.

Want to get in front of 13,000 AI enthusiasts? Work with me.

This newsletter is written & curated by Dario Chincha.

Affiliate disclosure: To cover the cost of my email software and the time I spend writing this newsletter, I sometimes link to products and other newsletters. Please assume these are affiliate links. If you choose to subscribe to a newsletter or buy a product through any of my links then THANK YOU – it will make it possible for me to continue to do this.