GROK
Grok 4 Fast targets high performance at low cost
xAI has released Grok 4 Fast, a model built to give strong reasoning at a lower cost.
It performs close to Grok 4 but uses far fewer tokens, making it cheaper and quicker to run.
The model handles both simple and complex tasks in one system, switching between short answers and detailed reasoning when needed.
It also supports a 2M token context window, advanced web and X search, and can use tools like code execution and browsing.
Independent checks show Grok 4 Fast is 98% cheaper than Grok 4 for the same benchmark results.
TL;DR
40% fewer tokens used on average, keeping accuracy high while cutting costs.
Stronger search tools across web and X, with better real-world results.
Wider access: free users can now try the latest Grok model.
AI that pays for itself
It also ranks first in LMArena’s Search Arena, beating larger rivals.
For users and developers, this means faster, lower-cost access to advanced reasoning.
Now available on grok.com, iOS, Android, OpenRouter, Vercel AI Gateway, and the xAI API, with separate reasoning and non-reasoning options.
Prices start at $0.20 per million input tokens.
Imagine doing the same job as Grok 4 but on a budget? That’s Grok 4 Fast.