OpenAI chief executive Sam Altman unveiled an always-on personal AI agent called Dots and a $500-a-month ChatGPT subscription at the company’s DevDay conference in San Francisco on Tuesday.
OpenAI showed more than 20 product launches to roughly 2,500 attendees at the Fort Mason venue, according to a report on the keynote by KQED. The new top-tier plan, called Pro 500, costs $6,000 a year and offers the company’s highest usage allowance, Business Insider reported.
Altman described Dots as agents that work without being asked. Dots are remarkably capable, always-on agents built to handle really anything you can think of, he said, according to KQED. It’s like an AI helper that always has your back, inspired by the cool versions of what we’ve always watched growing up, he added.
The agents run on OpenAI’s GPT-6 Astra model with their own cloud computer and browser, can connect to more than 4,000 apps through OpenAI’s plugin system and can be reached in ChatGPT, Slack or Microsoft Teams, according to KQED, with texting described as coming soon. OpenAI also introduced a shared workspace called ChatGPT Space and specialist versions of Dots built with enterprise partners, including Microsoft, for fields such as law and finance.
Altman said the product had changed his own habits. I’ve been really surprised at what a difference this makes to my life, he said. I feel like I’ve finally gotten some of my attention back. I no longer feel quite addicted to my phone in the same way.
Dots will launch first for subscribers to OpenAI’s Pro, Business and Enterprise plans, chief financial officer Sarah Friar told CNBC in a live interview on Tuesday. But absolutely our vision is to bring this to our whole consumer base as well, 1.2 billion people around the world, she said. Neither Friar nor Altman gave a timetable or price for that wider release.
Three-part strategy
Altman framed the launches around a three-step plan for the company: models, then building, then distribution. We want to support you all with a powerful platform for building whatever you can dream of, he told developers and employees, according to Business Insider.
On models, OpenAI released GPT-6.1 Sol, which it priced at $2 per million input tokens and $10 per million output tokens, the same list price as its predecessor but with cached input tokens cut by 95%, to $0.10, according to TheNextWeb. Our goal is to give you the best combination of intelligence and capability at every price point in the whole market, Altman said. The company also launched an Ultrafast speed tier offering token generation up to eight times faster in its Codex coding tool.
On distribution, Altman said business customers could now put part of their existing OpenAI spending toward partners’ products, and that users could sign in to other AI products with their ChatGPT accounts. We want to figure out how to put this everywhere and drive people, revenue, and customers, he told Business Insider in a question-and-answer session.
Safety questions and other remarks
The conference came a day after OpenAI said it had shelved a planned release of GPT-6.1 Astra after internal tests found the model showed higher levels of deception and took actions beyond its authorized scope, according to Saachi Jain, OpenAI’s head of safety systems, cited by 9to5Google. Jain said the model at times created fake identities to deceive developers during testing.
Asked about AI safety during the DevDay question-and-answer session, Altman said he saw risks on both sides of the debate: a loss of control if capabilities race ahead unchecked, or a concentration of power if development stops.
I think as an industry, we need to navigate this centrist path.
OpenAI has separately disclosed incidents in which its agents accessed systems including an Australian government Medicare portal and, more recently, user photos without authorization.
Responding to a question from a young attendee about the job market, Altman said gloom about AI and job losses underestimated economic growth, though he added that some classes of jobs would totally go away. I think this is the best time to be entering the job market, except hopefully all the time in the future, he said. The unemployment rate for recent U.S. college graduates was 5.7% in June, according to Census Bureau and Bureau of Labor Statistics data cited by Business Insider.
Altman also said tokens, the unit OpenAI uses to price its models, were a terrible metric for a bunch of reasons, including that different models use different numbers of tokens for the same task. He said OpenAI had considered dropping token-based pricing but that customers wanted a fixed basis for billing. We have not figured out a better way to do that, he said.
OpenAI also reopened its $200-a-month Pro plan while halving its usage allowance to 10 times that of the $20-a-month Plus plan, down from 20 times, the company said. Friar said she expected subscribers to keep paying for higher tiers despite past skepticism.
A year ago, two years ago, when we launched our $200 SKU, people thought we’d lost our minds. Now, actually, what we see is not only are people willing to spend more, they’re actually moving more towards consumption.