- Parameter count grows from 1.5 trillion (Grok 4.6) to 2.1 trillion
- Longer reinforcement learning and a greater focus on harder tasks aim to boost performance on coding, agentic work, and specialized-knowledge tasks
Why it matters: Pricing stays flat despite the upgrade, improving the cost-performance ratio for coding-agent use cases. Worth watching for teams evaluating AI coding assistants, since it raises capability without disrupting existing pricing plans.
KEY CONTEXTPricing is unchanged from Grok 4.6: $2 per million input tokens and $6 per million output tokens. Available via Cursor, Grok Build, and the xAI API.