Grok 4.7 is the newest and most capable model from Grok, designed to handle longer and more difficult tasks with ease. This latest iteration has been trained on a harder mix of tasks, weighted toward problems that take many hours to complete, and is better at verifying its own work and managing longer context.
One of the key improvements in Grok 4.7 is its ability to natively understand the Grok Bot harness, making it better at conversational tasks and general knowledge work. The model has also been refined to provide stronger cyber defense capabilities while maintaining low refusal rates for legitimate use.
In terms of performance, Grok 4.7 has shown impressive results on various benchmarks. On CursorBench 4.0, which stresses longer-running coding tasks, Grok 4.7 is at the frontier in price-performance. It also performs comparably to other frontier models on benchmarks like GDPval and AA Briefcase, which test professional knowledge work.
Safety and cybersecurity have been a major focus in the development of Grok 4.7. The model was built with an entirely new safeguard stack and has shown the highest safety on HackerBench v0.3, allowing only 3.3% of risky dual-use prompts through while rarely blocking legitimate security work. This balance of strong cyber defense and low refusal rates for legitimate use makes Grok 4.7 a powerful tool for a variety of applications.
Grok 4.7 is available today, priced starting at $2 per million input tokens and $6 per million output tokens. A fast variant with twice the output speed is also available at twice the price. Users can try Grok 4.7 for free in Grok Build or access it through the Grok API, third-party coding harnesses, and model routers and cloud platforms.
Este artigo foi escrito com a assistência de IA.
News Factory APP - notícias agênticas para impulsionar seu SEO e AEO.