OpenAI unveiled new API updates at its Dev Day on Monday, introducing GPT-5 Professional, its newest language mannequin, its new video technology mannequin Sora 2, and a smaller, cheaper voice mannequin.
The updates had been a part of a sequence of bulletins geared towards wooing builders to OpenAI’s ecosystem, together with the launch of an agent-building device and the potential to construct apps in ChatGPT.
The addition of GPT-5 Professional may enchantment to builders constructing functions in finance, authorized, and healthcare — industries that want “excessive accuracy and depth of reasoning,” per OpenAI CEO Sam Altman.
Altman additionally famous that voice capabilities can be important sooner or later because it rapidly turns into one of many main methods individuals use to work together with AI. To that finish, OpenAI is launching “gpt-realtime mini,” a smaller, cheaper voice mannequin in API that helps low-latency streaming interactions for audio and speech. The brand new mannequin is 70% cheaper than OpenAI’s previous advanced voice model however guarantees the “identical voice high quality and expressiveness.”
Lastly, creators concerned in OpenAI’s developer ecosystem can now faucet into Sora 2 in preview within the API. OpenAI launched Sora 2, its newest audio and video generator, final week alongside the Sora app, a TikTok competitor full of brief AI-generated movies. The Sora app permits customers to generate movies of themselves, mates, or something primarily based on a immediate, and share it by way of a TikTok-style algorithmic feed.
“[Developers] now have entry to the identical mannequin that powers Sora 2’s beautiful video outputs proper in your personal app,” Altman stated.
Sora 2 builds on its earlier technology with extra lifelike, bodily constant scenes with synchronized sound and better inventive management — from detailed digital camera path to stylized visuals.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
“For instance, you may take the iPhone view and immediate Sora to develop it right into a sweeping, cinematic vast shot,” Altman stated. “However some of the thrilling issues that we’ve been engaged on is how effectively this new mannequin pairs sound with visuals, not simply speech, however wealthy soundscapes, ambient audio, synchronized results which might be grounded in what you’re seeing.”
Sora 2 is pitched as a device for idea improvement, whether or not it’s a visible start line for an advert primarily based on the final vibe of a product, or a Mattel designer turning a sketch right into a toy idea — an instance Altman supplied at Dev Day that sheds mild on OpenAI’s take care of the Barbie-maker to convey generative AI into the toy-making pipeline.