OpenAI is preparing to release its next big language model (LLM), which powers the ChatGPT AI chatbot. The upcoming version, potentially named GPT-5, is expected to be significantly better than GPT-4.
According to a report from Business Insider, sources familiar with the matter state that OpenAI, led by Sam Altman, is on schedule to launch GPT-5 around mid-year, possibly during the summer.
Some companies have already seen demonstrations of the new model, comparing its improvements to the ChatGPT tool.
A CEO who previewed a version of GPT-5 commented positively, stating it’s “really good, like significantly better.” OpenAI showcased the new model using specific examples and data relevant to the CEO’s company. The CEO also mentioned some unreleased features, such as the ability to summon AI agents for autonomous tasks.
OpenAI still training GPT-5
OpenAI is currently still training GPT-5 and will conduct internal safety tests afterward. Following that, the model will undergo “red teaming,” where both employees and selected outsiders challenge the model to identify any issues before its public release.
There’s no set timeframe for completing safety testing, and the release date for GPT-5 remains unknown. OpenAI generates revenue primarily by selling its AI models to business clients.
Last year, OpenAI introduced GPT-4 in March and later enhanced its capabilities with GPT 4-Turbo in November. This updated model is more capable and has knowledge of world events up to April 2023. It also addresses issues like “laziness,” where the model occasionally declines to respond to prompts.
Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook.
