OpenAI is set to introduce GPT-4o, an evolution of the GPT-4 model fueling its flagship product, ChatGPT. Mira Murati, OpenAI's CTO, highlighted the enhanced speed and expanded capabilities across text, vision, and audio during a livestream announcement. Notably, GPT-4o will be freely accessible to all users, with paid users enjoying increased capacity limits up to five times that of free users, as per Murati.
In a recent blog post, OpenAI outlined the gradual rollout of GPT-4o's features, with text and image capabilities becoming available in ChatGPT starting today. CEO Sam Altman emphasized the model's native multimodal capabilities, enabling it to comprehend and generate content through voice, text, or images seamlessly. Altman also revealed that developers keen on exploring GPT-4o will have access to an API that boasts faster speeds and more affordable pricing compared to GPT-4 Turbo.
Exciting updates are on the horizon for ChatGPT’s voice mode with the introduction of the new model. Transforming into a Her-like voice assistant, the app will seamlessly engage with users, offering real-time responses while intelligently perceiving its environment. In contrast, the existing voice mode operates within more confined parameters, addressing one prompt sequentially and relying solely on auditory input.
In a reflective blog post following the livestream event, Altman delved into OpenAI's evolving trajectory. Initially driven by a vision to bring manifold benefits to the world, the company has undergone a notable shift in focus. Critics have pointed out OpenAI's reluctance to open-source its advanced AI models. Altman appears to acknowledge this criticism, indicating a new direction. Instead of solely creating AI for broad benefits, the emphasis now lies in providing access to these models through paid APIs. The intention is for developers and third parties to utilize these tools creatively, fostering innovation and generating widespread benefits. This pivot suggests a strategic move towards a collaborative approach in advancing AI technology.
During a livestream announcement on Monday, OpenAI CTO Mira Murati unveiled significant updates to ChatGPT. Now, all users can enjoy these features for free, with paid users benefiting from up to five times the capacity limits. This move democratizes access to advanced functionalities previously exclusive to paid subscriptions. Free users now have access to web searches, varied voice options, and the ability to store and recall information. Moreover, OpenAI is gradually rolling out GPT-40's enhanced text and image capabilities to paying ChatGPT Plus and Team users. Enterprise users can expect these upgrades soon. Additionally, ChatGPT Plus users will soon experience the new version of the voice mode assistant.
- GPT-40's expanded capabilities now include vision functionalities, enabling it to interpret desktop screenshots directly from Mac devices.
- Mobile app integration facilitates interaction with GPT-40 through an iPhone app, with compatibility for Windows coming soon. Users can upload videos and screenshots for processing.
- In demonstrations, OpenAI showcased GPT-40's enhanced human-like qualities. During real-time interactions, the voiced ChatGPT engages in banter and responds with humor.
- Unlike its predecessors, GPT-40 allows seamless back-and-forth conversations without waiting for the model to complete its responses.
- With harmonized speech synthesis, GPT-40 can produce various voices and blend them for a more natural conversational flow.
- GPT-40 offers sophisticated interactions, including translations and "normal" conversational exchanges, showcasing the advancements expected from GPT-4 level technology.
