OpenAI Launches GPT-4o, an Upgraded Iteration of GPT-4
OpenAI has announced the launch of GPT-4o, an updated version of the GPT-4 model that powers its flagship product, ChatGPT. According to OpenAI CTO Mira Murati, the new model is significantly faster and boasts improved capabilities across text, vision, and audio. The update will be free for all users, with paid users enjoying up to five times the capacity limits of free users.
Iterative Rollout and Multimodal Capabilities
In a blog post, OpenAI stated that GPT-4o’s capabilities will be rolled out iteratively, with text and image capabilities being the first to be introduced in ChatGPT. OpenAI CEO Sam Altman described the model as “natively multimodal,” allowing it to generate content or understand commands in voice, text, or images. Developers interested in experimenting with GPT-4o will have access to the API, which is half the price and twice as fast as GPT-4 Turbo.
Enhanced Voice Mode Features
ChatGPT’s voice mode will receive new features as part of the GPT-4o update. The app will be able to function as a voice assistant, similar to the AI in the movie “Her,” responding in real-time and observing the user’s surroundings. This is a significant improvement over the current voice mode, which is limited to responding to one prompt at a time and working only with what it can hear.
OpenAI’s Shifting Vision
In a blog post following the livestream event, Altman reflected on OpenAI’s trajectory, acknowledging that the company’s original vision of creating benefits for the world had shifted. Despite criticism for not open-sourcing its advanced AI models, Altman suggested that the company’s focus has changed to making those models available to developers through paid APIs, allowing third parties to create innovative applications that benefit everyone.
Timing and Speculation
The GPT-4o launch comes just ahead of Google I/O, the tech giant’s flagship conference, where various AI products from the Gemini team are expected to be unveiled. Prior to the announcement, there were conflicting reports about what OpenAI would be launching, with predictions ranging from an AI search engine to rival Google and Perplexity, a voice assistant integrated into GPT-4, or an entirely new and improved model, GPT-5.
Model evaluations
As measured on traditional benchmarks, GPT-4o achieves GPT-4 Turbo-level performance on text, reasoning, and coding intelligence while setting new high watermarks on multilingual, audio, and vision capabilities.
Improved Reasoning – GPT-4o sets a new high score of 88.7% on 0-shot COT MMLU (general knowledge questions). All these evals were gathered with our new simple evals(opens in a new window) library. In addition, on the traditional 5-shot no-CoT MMLU, GPT-4o sets a new high score of 87.2%. (Note: Llama3 400b(opens in a new window) is still training)
Need help integrating AI into your business?
Brain Buzz Marketing is an expert in AI technology. We help businesses use advanced AI models like Chat GPT, Co-Pilot, Claude, Gemini, Perplexity, Mistral, DALL- E3, Leonardo AI, Stable Diffusion, Mid Journey, and Adobe Fire Fly.
Contact Brain Buzz Marketing to set up a free AI consultation.
