Elon Musk Teases Grok 4.5, Says New Model Matches Top AI Rivals

Elon Musk has announced that Grok 4.5, the next version of xAI’s chatbot, has entered private beta testing at SpaceX and Tesla. According to Musk, early internal evaluations suggest the model performs close to, and in some cases exceeds, Anthropic’s Claude Opus, although these claims have not been independently verified.

In a series of posts on X, Musk said Grok 4.5 is built on xAI’s new 1.5-trillion-parameter V9 foundation model and has received additional training using data from the AI coding platform Cursor. He added that reinforcement learning continues to improve the model and praised the teams involved in its development.

Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.

RL is continuing to significantly improve the model, and the Grok Build…

— Elon Musk (@elonmusk) June 28, 2026

Musk also outlined an ambitious roadmap, saying, “The new AI models trained from scratch will be released every month for the rest of the year, reflecting a much faster development cycle than previous versions.”

Clarifying his earlier comments, Musk said he was not suggesting that the V9 foundation model would dramatically outperform every competing AI system. Instead, he described it as “a solid workhorse in the same league as Opus”, with future improvements expected through continuous model training and optimisation.

He also said Grok 4.5 would represent a significant upgrade over Grok 4.3, which was based on the older 500-billion-parameter V8 model trained in December. According to Musk, the V8 model has several fundamental limitations that the new version aims to address.

Beyond the model itself, Musk said xAI is accelerating development by improving its software infrastructure. He revealed that engineers from SpaceX’s Starlink and Starship programmes have shifted much of their focus to the company’s AI efforts, helping speed up model development.

The latest announcement comes as competition in artificial intelligence continues to intensify, with companies including OpenAI, Anthropic, Google and xAI racing to build more capable large language models. If xAI delivers on its plan to release new models every month, the company could significantly increase the pace of competition in the AI industry over the coming months.