Meet Moshi, a new AI chatbot with GPT-4o-like abilities

French AI company Kyutai has developed a new AI chatbot called "Moshi" with features similar to ChatGPT's now-delayed "enhanced voice mode" GPT-4o. Moshi understands the tone of your voice and interprets it. It can also be used offline.

Based on a 7B parameter large language model (LLM) called Helium, the chatbot is currently available to everyone and can speak with different accents and 70 different emotional and speech styles. Moshi can also handle two sound currents at the same time, which means it can listen and talk at the same time.

AI Chatbot, named after the Japanese way of answering a phone call, has a response time of only 200 milliseconds, making it faster than the GPT-4OS advanced voting state, which typically takes anywhere between 232 to 320 milliseconds. Kyutai, a French AI company, has developed a new AI-powered chatbot called "Moshi" that offers features similar to ChatGPT's now-delayed 'Advanced Voice Mode' GPT-4o. Moshi can understand your tone of voice and interpret it. It also works offline.

A chatbot based on the 7B-parameter Large Language Model (LLM) called Helium is currently available to everyone and can speak in a variety of accents and 70 different emotions and speech styles. Moshi can also handle two audio streams simultaneously, meaning it can listen and talk at the same time.

The AI ​​chatbot, named after the Japanese language for answering a phone call, has a response time of just 200 milliseconds, making it faster than GPT-4o's advanced voice mode, which typically takes between 232 and 320 milliseconds. It turns out that Kyutai is also developing an AI-powered audio recognition, watermarking, and signature tracking system that will eventually integrate with Moshi. While this may not be the ChatGPT competitor we were hoping for, it's certainly a step forward in developing open source models that can work offline.