Elon Musk-backed xAI has launched Grok 4 Fast, an optimized version of its Grok 4 artificial intelligence system, aimed at delivering high performance at lower operational costs. The new model is built to cater to both enterprise and consumer applications, combining reasoning and non-reasoning capabilities within a single, unified framework. This integration reduces the need for separate models, cutting latency and simplifying deployment.
According to xAI, Grok 4 Fast consumes 40% fewer “thinking tokens” on average compared to its predecessor, without compromising accuracy. Independent analysis by Artificial Analysis confirmed a significant improvement in cost efficiency, estimating that Grok 4 Fast requires up to 98% less spending to achieve results comparable to Grok 4.
Benchmark results highlight the model’s strong capabilities. Grok 4 Fast scored 85.7% on the GPQA Diamond test, 92% on AIME 2025, and 93.3% on HMMT 2025, closely mirroring Grok 4’s performance. The model also demonstrated superior performance in code execution and search-based benchmarks, including a 95% pass rate on SimpleQA and 74% on X Bench Deepsearch.
One of Grok 4 Fast’s standout features is its 2 million token context window, allowing the AI to efficiently handle large-scale inputs. Reinforcement learning techniques were applied during training to further optimize its reasoning efficiency and speed.
xAI has made Grok 4 Fast widely accessible. Users can access it through grok.com, as well as on iOS and Android apps. It is also integrated with platforms like OpenRouter, Vercel AI Gateway, and the xAI API, with free access available for a limited time on OpenRouter and Vercel.
The model comes in two variants — grok-4-fast-reasoning and grok-4-fast-non-reasoning — both supporting the 2 million token limit. Pricing starts at $0.20 per million input tokens, scaling according to usage, making it an attractive option for businesses and developers seeking high-quality AI performance at a lower cost.
Grok 4 Fast marks xAI’s latest step toward more efficient, scalable, and versatile AI tools, reflecting ongoing efforts to make advanced AI more accessible and affordable across industries.
