Opening article
OpenAI has released GPT‑4o, a multimodal model that can process and generate text, audio, image, and video in real time. It responds to audio in about 320 ms, matching GPT‑4 Turbo on text and code while outperforming it on vision and audio tasks. GPT‑4o is available in ChatGPT’s free tier and Plus plan, and through the API as a text‑and‑vision model that is twice as fast, half the price, and five times the rate limit of GPT‑4 Turbo.
GPT‑4o (“o” for “omni”) is a step towards much more natural human-computer interaction—it accepts as input any combination of text, audio, image, and video and generates any combination of text, audio, and image outputs. It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time (opens in a new window) in a conversation. It matches GPT‑4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50% cheaper in the API. GPT‑4o is especially better at vision and audio understanding compared to existing models.
Prior to GPT‑4o, you could use Voice Mode to talk to ChatGPT with latencies of 2.8 seconds (GPT‑3.5) and 5.4 seconds (GPT‑4) on average. To achieve this, Voice Mode is a pipeline of three separate models: one simple model transcribes audio to text, GPT‑3.5 or GPT‑4 takes in text and outputs text, and a third simple model converts that text back to audio. This process means that the main source of intelligence, GPT‑4, loses a lot of information—it can’t directly observe tone, multiple speakers, or background noises, and it can’t output laughter, singing, or express emotion.
With GPT‑4o, we trained a single new model end-to-end across text, vision, and audio, meaning that all inputs and outputs are processed by the same neural network. Because GPT‑4o is our first model combining all of these modalities, we are still just scratching the surface of exploring what the model can do and its limitations.
Developers can also now access GPT‑4o in the API as a text and vision model. GPT‑4o is 2x faster, half the price, and has 5x higher rate limits compared to GPT‑4 Turbo. We plan to launch support for GPT‑4o's new audio and video capabilities to a small group of trusted partners in the API in the coming weeks.
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.
These 20 languages were chosen as representative of the new tokenizer's compression across different language families
હેલો, મારું નામ જીપીટી-4o છે. હું એક નવા પ્રકારનું ભાષા મોડલ છું. તમને મળીને સારું લાગ્યું!
The integration of real‑time audio and vision into a single model simplifies developer workflows and enables richer, more natural user interactions, potentially accelerating adoption of multimodal AI in consumer and enterprise products.