Medium Foundation Models · 1 min read
Grok 4: xAI puts reasoning at the center and introduces multi-agent 'Grok 4 Heavy'
In one sentence xAI launches Grok 4 and Grok 4 Heavy (variant running multiple parallel instances, like o1-pro). SuperGrok Heavy tier at $300/month. High but contested benchmark numbers.
Reading level
Five months after Grok 3, xAI ships Grok 4 in a livestream. Two new things:
- Grok 4: the base model, with reasoning always on (no longer a toggle);
- Grok 4 Heavy: a "boosted" version running multiple model instances in parallel, having them reason separately, then synthesizing the best answer. Similar to OpenAI o1-pro.
The company claims performance above all competitors on benchmarks, particularly "Humanity's Last Exam" — a test of extremely hard questions designed to push models to their limit.
Pricing: SuperGrok ($30/month) for regular Grok 4, SuperGrok Heavy ($300/month) for the Heavy version. Available via developer API too. Community greets benchmark numbers with skepticism but acknowledges it competes with o3 and Claude 4.
Companies
xAI
Tools
Grok 4, Grok 4 Heavy
Tags
xAIGrok 4ReasoningElon Musk
Sources