Elon Musk teased the official Grok CLI for developers looking to beat Anthropic Claude in the agent coding race.
The announcement was made by xAI’s founders in response to user posts about their dissatisfaction with Anthropic, a popular terminal-based AI coding tool that handles file editing, shell commands, and complex codebases directly within a developer’s workflow.
“Hey @elonmusk, if you hate Anthropic so much, just make a Grok CLI and I’ll cancel my Anthropic subscription today,” the user wrote in the post.
Elon Musk replied, “Soon.” No details regarding dates, prices, or features were mentioned.
Community versions such as superagent-ai/grok-cli already exist using xAI’s API, and Musk’s tease hints at competing official versions in the agent coding space.
This big announcement comes a day after Elon Musk accused Anthropic of stealing training data on a large scale.
“Anthropic was guilty of mass theft of training data and had to pay a multi-billion dollar settlement for that theft. These are just the facts,” SpaceX’s CEO wrote to X.
Prior to these claims, Anthropic accused Chinese AI companies such as DeepSeek, Moonshot AI, and MiniMax of using Claude for unauthorized model training and shared disturbing findings.
According to the US-based artificial intelligence company, these startups have rapidly improved the performance of low-scale models using a technique called “distillation,” which extracts and leverages output from more powerful AI systems.
“These labs created over 24,000 fraudulent accounts and generated over 16 million interactions with Claude, unlocking the ability to train and improve their own models,” Anthropic posted.
