OpenAI introduced a slew of new products at its DevDay event on Tuesday, including agents that can work in the background, a cheaper AI model, and new ways to use ChatGPT across other apps.
The biggest new product is dots, which OpenAI bills as an "always-on" AI agent similar to Grok Bot or Meta's Muse that can keep working on a task after a user steps away. The deck showcased at DevDay says a dot can use connected apps, its own browser, and a cloud computer for tasks like preparing code changes, updating a project plan, or tracking customer feedback.
OpenAI also launched GPT-6.1 Sol, which it says improves on GPT-6 Sol in coding and other technical work. The company says GPT-6.1 Sol's performance comes close to its top-end Astra model on some tasks, but at one-fifth of Astra's input and output token prices. It also pitched a new Ultrafast option promising faster responses for users willing to pay more, and a $500 monthly Pro plan with its highest usage allowances.
Other products and offerings focused more on cross-platform and cooperative work. ChatGPT Spaces allows teams a place to share documents and projects, while "Sign in with ChatGPT" will let customers use a range of third-party tools like Notion and Vercel with their ChatGPT allowances.
The company also revealed updated user numbers, touting a massive 1.2 billion weekly ChatGPT users. That's an increase of roughly 300 million WAUs in the past seven months alone.
This comes as OpenAI is reportedly seeking $30 billion in funding at a $1.4 trillion pre-money valuation. Asked about those reports on CNBC, CFO Sara Friar declined to comment.
She said the company was well capitalized but needed to keep investing in compute power as demand continues to rise.
Friar also declined to put a date on a potential initial public offering. She called going public "a milestone in the journey" and said OpenAI wants to do it "when the time is right for our business."
CEO Sam Altman made a similar point earlier in the day in a separate interview.
“I don't have a particular timeline in mind," he said, adding that an IPO would make more sense after the company has run "a few safety cases" and is better prepared to manage the next level of AI.