In artificial intelligence and natural language processing (NLP), a token is a single unit of text used by language models to process and generate information. Tokens can represent words, subwords, characters, or even punctuation marks, depending on how the model is designed. During training and inference, AI models break text into tokens to understand structure, meaning, and context. For example, a short sentence might be divided into several tokens that the model analyzes sequentially to predict the next word or generate coherent text. The concept of tokens is essential for managing input length, optimizing performance, and calculating costs in large language models.
Chinese generative AI startups MiniMax and Zhipu AI have disclosed their financials ahead of potential Hong Kong listings, highlighting modest revenues and mounting losses compared with U.S. peers.
Amazon is reportedly in early talks to invest up to $10 billion in OpenAI in a deal tied to the use of its AI chips. The discussions reflect a broader trend of circular partnerships shaping the AI infrastructure market.
Zoom introduced AI Companion 3.0, a federated AI solution integrating internal and third-party models to streamline meetings, documents, and personal workflows. The update adds agentic AI features, a web interface, and enhanced productivity tools.
Nvidia is reportedly planning to increase production of its H200 AI training chips following approval to sell them in China.
OpenAI has released GPT-5.2, a new model series designed to improve professional knowledge work with major gains in reasoning, coding, long-context analysis, and vision. The update brings enhanced performance to ChatGPT and the API across Instant, Thinking, and Pro variants.
Disney and OpenAI announced a three-year licensing deal allowing Sora to generate fan-inspired short videos featuring over 200 Disney, Marvel, Pixar, and Star Wars characters.
OpenAI faces a wrongful death lawsuit alleging ChatGPT reinforced a man’s violent delusions, contributing to the killing of his mother in Connecticut. The case marks the first time an AI platform has been accused of involvement in a murder.
Google introduced the Titans architecture and the MIRAS framework to enable AI models to handle massive contexts and update their internal memory while running, improving performance in long-sequence tasks.
ChatGPT now reaches more than 800 million weekly users, creating a powerful adoption flywheel that is speeding the transition from AI experimentation to full-scale enterprise deployment, according to OpenAI’s new 2025 report.
OpenAI buys Neptune for under $400 million in stock, gaining advanced AI model tracking tools and strengthening capabilities in large-scale AI development.