xAI's economical Grok 4 Fast

A new hyper-efficient reasoning model called Grok 4 Fast was introduced by xAI. It offers near-frontier performance and peak speed at a quarter of the compute cost of Grok 4. The specifics: Grok 4 Fast uses 40% less thinking tokens on average, which translates in a 98% price decrease, but still produces outcomes that are comparable to Grok 4. According to benchmarks, it outperforms Claude 4.1 Opus and Gemini 2.5 Pro, scoring 92% on AIME 2025 (math) and 85.7% on GPQA Diamond (science). Additionally, the model outperformed the larger Grok 4 on coding benchmarks and achieved the top spot in LMArena's Search Arena. In addition to native tool integration for web surfing and code execution, Grok 4 Fast provides a 2M token context. Despite significant cost reductions, Grok 4 Fast is now competitive with the best models worldwide because to xAI's incredible cost-efficiency improvements with this latest release. This model is a part of the trend that reflects the impending reality when leaders like Sam Altman talk about "intelligence too cheap to meter."