OpenAI has unveiled GPT-4 Turbo, an enhanced version of its groundbreaking AI model, at its inaugural in-person event. This advancement not only offers users the capability to develop personalized ChatGPT chatbots but also significantly reduces operational costs for businesses and developers utilizing its software. Demonstrating rapid growth, OpenAI reported approximately 100 million weekly active ChatGPT users, with adoption by over 92% of Fortune 500 companies, expanding from various sectors like finance, law, and education.
Nadella Onstage
“The systems that are needed as you aggressively push forward on your road map require us to be on the top of our game, and we intend fully to commit ourselves fully to making sure you all… have not only the best systems for training and inference, but also the most compute,” said Microsoft CEO Satya Nadella to OpenAI CEO Sam Altman as they were onstage together. “That’s the way we’re going to make progress.”
Following Microsoft’s substantial investment of $10 billion earlier this year, OpenAI has continued to scale, reaching valuations up to $29 billion with contributions from leading investors. Discussions of an $80 billion valuation have been on the table as well.
During his Monday address for the OpenAI DevDay event, Altman highlighted that the day’s updates were a direct result of engaging with developers over the past year to understand their requirements. He also touched upon the future developments concerning GPT-5.
No Timeline
“We want to do it, but we don’t have a timeline,” said Altman. “We are just as annoyed as all of you, probably more, that GPT’s knowledge about the world ended in 2021.”
The new GPT-4 Turbo model boasts updated information up to April 2023, accepts inputs as extensive as 300 pages, and introduces DALL-E 3 for AI-generated imagery and enhanced text-to-speech capabilities with multiple voice options. Available for developer preview, GPT-4 Turbo is slated for wider release soon.
Additionally, OpenAI has announced cost reductions for its developer services, with input tokens priced 70% lower and output tokens 50% cheaper than previous rates. The company is also launching the GPT Store, allowing developers to monetize their custom AI chatbots based on usage.
“Anyone can easily build their own GPT — no coding is required,” wrote the company in a release. “You can make them for yourself, just for your company’s internal use, or for everyone. Creating one is as easy as starting a conversation, giving it instructions and extra knowledge, and picking what it can do, like searching the web, making images or analyzing data.”
The update consolidates various AI tools from OpenAI into a singular user-friendly platform, integrating DALL-E image generation, browsing, and unique features like PDF search — a capability until now only offered by Anthropic’s Claude. This streamlining aims to enhance the user experience and lower the barrier to entry for utilizing OpenAI’s AI suite.
