Grok Lite briefly sent gibberish to users
Archive item — written before sources were shown.
A subset of Grok.com users got nonsensical output this week; xAI called it a rare temporary generation glitch and said starting a new chat usually clears it.
Some Grok.com users, on the free Grok Lite tier specifically, started getting nonsensical responses this week, strings of disconnected words or stray links to reinforcement-learning research sites instead of coherent answers; the Grok account on X wasn’t affected. xAI’s official account acknowledged it as “a rare temporary generation glitch” that usually clears with a fresh chat or regenerated response, and the company’s status page showed all services as operational. TechCrunch couldn’t reproduce the issue directly, suggesting it’s hitting a limited slice of users rather than the full Grok Lite population.
- 01Grok keeps sending gibberish responses to userstechcrunch.com · primary reporting
