2 min read
Just months after launching Grok 4 and facing backlash over a highly controversial antisemitic meltdown from its chatbot, xAI is shifting gears with a new release. Elon Musk’s AI company has introduced Grok 4 Fast, a model designed to deliver quicker results while using fewer resources.
According to xAI, Grok 4 Fast achieves performance comparable to Grok 4 but consumes 40 percent fewer thinking tokens on average. The company also claims the model cuts costs dramatically, reducing the price of achieving Grok 4-level performance on frontier benchmarks by 98 percent. That applies to tasks ranging from writing code to browsing the web for instant answers.
The update also reflects a change in architecture. Like OpenAI’s GPT-5 approach, Grok 4 Fast blends two modes: a reasoning model for complex problems and a non-reasoning model for lightweight queries. This structure allows the system to switch seamlessly depending on the nature of the request.
In tests, on a benchmarking platform that compares AI models side by side, Grok 4 Fast ranked first in search-related tasks and eighth in text generation. The model is already live for all users, including free-tier accounts, across web, iOS, and Android.
