ChatGPT adds new GPT-4o (S2S) Alpha model for desktop alongside GPT 4 and GPT-4o
As we reported previously, OpenAI is about to launch a new advanced Voice Mode for ChatGPT with a new Alpha model. GPT-4o, their new flagship model, will reason across audio, vision, and text in real time, which will be rolling out in the coming weeks. Looks like the new Alpha model is right around the corner as some desktop users are seeing a new GPT-4o (S2S) Alpha model.
ChatGPT Plus users are seeing a new model popup alongside the GPT 4 and GPT 4-omni — called ChatGPT-4o (S2S) which stands for Sequence-to-Sequence.
However, the new advanced Voice functionality is not yet implemented.
What is GPT-4o S2S model?
ChatGPT’s “S2S” refers to the model’s architecture, which is optimized for tasks like language translation, summarization, and text generation.
However, it may not necessarily be for voice capabilities. Voice functionality requires technologies like Text-to-Speech (TTS) and Automatic Speech Recognition (ASR), which are different from the text-based capabilities of ChatGPT.
The sequence-to-sequence architecture is designed for tasks involving pairs of input and output sequences, such as conversational exchanges. This model enables the AI to generate more coherent and contextually appropriate responses.
With S2S the AI draws from a more extensive, less filtered database for a more diverse range of responses.
It is expected that the new model can understand and generate responses that are more contextually relevant and natural, making interactions feel more human-like.
The New Voice Mode is still coming soon with new GPT-4o capabilities in an alpha within ChatGPT Plus. Maybe this is the first step in that direction.
