What's Happening in OpenAI?

GPT-4o Release — GPT-4o is the new flagship model from OpenAI. OpenAI launched the GPT-4o model on the 13th of May 2024 with features like lower latency and multimodal capabilities at half the price of GPT-4 Turbo. This new model supports text, voice, and image inputs, providing more versatile and efficient interactions. One of its standout features is its ability to respond to audio input in just 232 milliseconds, achieving human-level response times. This model is available in preview on the Azure OpenAI Service.

Whisper Automatic Speech Recognition (ASR) Model — OpenAI announced updates to its ASR model in the month of May 2024. Whisper large-v3 is the latest iteration that showcases improved performance across various languages. This model was trained on a diverse dataset of 1 million hours of weakly labelled audio and 4 million hours of pseudolabelled audio, collected using the previous large-v2 model. The updates have led to a 10% to 20% reduction in error rates compared to the previous version, enhancing its robustness in multilingual speech recognition tasks. It is available on GitHub under a permissive license, enabling developers to integrate it into various applications. The new version enhances recognition of…