xAI has officially rolled out its new lightweight flagship model, Grok4Fast. This innovative model not only cuts down the computational load by a remarkable 40% but also has the potential to reduce single-task operational costs by as much as 98%.

Equipped with a context window at the 2-million-token level, Grok4Fast is well-suited for handling intricate analyses and time-critical tasks. Its performance has been put to the test in rigorous benchmark evaluations. In the GPQA Diamond benchmark test, the model achieved an impressive score of 85.7%. Moreover, during the AIME2025 professional evaluation, it excelled with a score of 92.0%.

When it comes to commercialization, xAI has adopted a tiered pricing strategy. The cost per million tokens varies, ranging from a mere $0.05 to $1.00, offering flexibility to different user groups. What's more, users have the opportunity to try out Grok4Fast free of charge through the OpenRouter and Vercel platforms, allowing them to experience its capabilities firsthand.