At Monday, November 6’s DevDay, OpenAI announced plenty of exciting updates for developers and ChatGPT users. CEO Sam Altman took the stage to share details of new products, pricing, how to make GPT your own, and then monetize it. As Altman says, “If you give people better tools, they do better things.” Keep reading to learn how to use new GPT products to do more on your behalf and make your own GPTs available to others.

You can watch CEO Sam Altman’s entire keynote here, but if you’re looking for announcements and high-level takeaways, we have you covered below!

What is GPT-4 Turbo and how much does it cost?

• OpenAI’s newest product, GPT-4 Turbo is now available via an API preview. It’s a faster and smarter model than GPT-4 with increased context.
• GPT-4 Turbo has the longest-ever context length: 128k and higher rate limits: 2x tokens per minute
• GPT-4 Turbo is available at a 3x cheaper price for input tokens and a 2x cheaper price for output tokens compared to GPT-4
• GPT-4 Turbo accepts the equivalent of more than 300 pages of text in a single prompt
• Turbo can accept images as inputs, allowing you to generate captions, analyze images in detail, and understand documents with figures
• Vision support will be added to the main GPT-4 Turbo model 

Did OpenAI introduce new pricing?

• Exciting news! GPT-4 Turbo is approximately one-third the input cost and half the output price of the GPT-4 model. Other ChatGPT models have adjusted pricing now too.
• GPT-4 Turbo input tokens are $0.01 and output tokens are $0.03
• GPT-3.5 Turbo input tokens are 3x cheaper than the previous model at $0.001 and output tokens are 2x cheaper at $0.002

What is ChatGPT’s new knowledge cutoff date? 

• GPT-4 Turbo has knowledge up until April 2023! The cutoff date for the GPT-4 model is January 2022.

What are GPTs?

Altman also introduced GPTs, which he referred to as the number-one developer request. These are tailored versions of ChatGPT you create for a specific purpose, whether professional or personal. Use GPTs to build a customized version of GPT for almost anything, then publish it for others to use. Combining instructions, expanded knowledge, and action, GPTs can work better for you and offer more control.

Anyone can easily build their own GPT, and there’s no coding required. Altman says it’s as simple as starting a conversation, providing instructions and additional context, and specifying what it can do, like browsing the web, analyzing data, or creating imagery.

Later in November, OpenAI will launch the GPT store. You can list your GPT there, and the best and most popular GPTs will be featured. Even better, the people who build the most useful and used GPTs will receive a portion of the revenue earned! Custom GPTs have to follow OpenAI policies before being made available.

What other updates did OpenAI announce at DevDay?

• New multimodal capabilities in the platform, including vision, image creation (DALL·E 3), and text-to-speech (TTS)
• New version of GPT-3.5 Turbo that supports a 16K context window 
• More control with JSON, functions, and viewing logs
• ChatGPT Plus and Enterprise users can integrate DALL·E 3 directly into their apps and products through OpenAI’s Images API by specifying dall-e-3 as the model
• Generate human-quality speech from text via OpenAI’s text-to-speech API, which offers six preset voices and two model variants
• OpenAI is launching a Custom Models program, giving select organizations the opportunity to work with OpenAI researchers to train custom GPT-4 to their domain
• Expanded partnership with Microsoft

These new features are available to OpenAI customers now! Follow Brighter Future AI on YouTube for daily AI updates.