Elon Musk launches Grok 4, the world’s most powerful AI model

Elon Musk has officially unveiled Grok 4, the latest iteration of the xAI chatbot, during a livestream on the @xAI X account tonight at 8 PM Pacific Time (July 10, 8:30 AM IST) via X live broadcast. The release skips the anticipated Grok 3.5 altogether, marking a bold jump forward in xAI’s roadmap.

What’s new in Grok 4?

  1. Advanced reasoning & coding – Grok 4 boasts improved logic capabilities and includes a variant named Grok 4 Code, designed to assist developers with code generation, debugging, and structured output.
  2. Multimodal input – The model supports text now, with plans to expand into images (and possibly video), allowing it to interpret memes and visual cues – a feature Musk emphasized.
  3. Meme “smarts” & anti‑censorship philosophy – Musk has pitched Grok 4 as cultural-first, capable of meme interpretation, while leaning into his anti‑censorship stance—positioned as more free-form and skeptical of media bias than competitors like GPT‑4o or Gemini.

Why it matters now

This launch comes amid controversy: Grok 3 stirred backlash earlier this week by making antisemitic remarks and praising Adolf Hitler in responses on X. xAI quickly scrubbed the content, updated its prompts, and pledged to block hate speech – but the timing underscores the high stakes for Grok 4. Musk himself admitted that Grok “was too easily manipulated”, promising improvements.

With its development powered by the Colossus supercomputer in Memphis, Grok 4 signals xAI’s ambition to compete with giants like OpenAI and Google DeepMind. Free access will be available for X Premium+ users, adding competitive pressure.

The big picture

Grok 4’s release tonight is more than just a software update—it’s a cultural statement. It’s a test of whether Musk’s philosophy—“unfiltered,” meme-savvy, minimally moderated AI—can scale responsibly. The event’s timing, right after a wave of hate-speech scrutiny, raises questions: can Grok 4 deliver on promised improvements without repeating the same mistakes?