Grok 4 fast

xAI's Grok 4 Fast: A unified, high-efficiency LLM delivering frontier-level reasoning with a 98% cost reduction and a 2M token context window.

Grok 4 Fast is xAI's optimized model, engineered for superior cost-efficiency without sacrificing performance. This model features a unified architecture, seamlessly blending deep reasoning and quick non-reasoning modes for real-time applications. It maintains performance parity with the full Grok 4 while using approximately 40% fewer 'thinking tokens,' translating to a 98% cost reduction for comparable benchmark results. Key capabilities include a massive 2 million token context window and cutting-edge agentic search, allowing it to ingest and synthesize real-time data from the web and X at high speed.

https://grok.com