OpenAI ramps up developer push with more powerful models in its API

During its recent Dev Day, OpenAI announced significant enhancements to its API, showcasing the introduction of GPT-5 Pro, its latest language model, alongside a novel video generation model named Sora 2 and an affordable voice model. These updates aim to attract developers to the OpenAI ecosystem and include the launch of an innovative agent-building tool as well as the capability to create applications directly within ChatGPT. The introduction of GPT-5 Pro is particularly noteworthy for developers focusing on sectors such as finance, healthcare, and legal, where precision and deep reasoning are essential. OpenAI CEO Sam Altman emphasized the growing importance of voice interactions, hinting that they could become a dominant method for engaging with AI in the near future. To facilitate this transition, OpenAI is rolling out "gpt-realtime mini," a compact and cost-effective voice model that ensures low-latency streaming for audio and speech interactions, priced 70% lower than its predecessor while maintaining high-quality voice outputs. Additionally, developers can now access Sora 2 in preview through the API. Released alongside the Sora app—an emerging competitor to TikTok that features a variety of AI-generated short videos—Sora 2 enhances the audio and video generation experience. Users can create personalized videos based on prompts, sharing them through a TikTok-like algorithmic feed. Altman noted that developers are now empowered with the same model that drives Sora 2’s impressive video capabilities within their own applications. Sora 2 represents a leap forward from its predecessor, offering more realistic scenes, synchronized sound, and enhanced creative control, including intricate camera direction and stylized visuals. For instance, users can prompt Sora to transform a standard iPhone view into a dramatic cinematic wide shot. One of the standout features of this new model is its ability to seamlessly integrate sound with visuals, creating immersive experiences that go beyond speech to include rich soundscapes and ambient audio. This tool is envisioned as a resource for concept development, aiding industries from advertising to toy design, as highlighted by Altman's collaboration with Mattel to integrate generative AI into the toy-making process.