xAI ships Grok 4.7: a bigger model at the same price
xAI's new 2.1-trillion-parameter model is 40% bigger than Grok 4.6 at the same price. It placed second on key benchmarks.
The model
xAI released Grok 4.7 on September 21. It has 2.1 trillion parameters, 40% more than the 1.5 trillion in Grok 4.6, and a 500k-token context window. Pricing is $2 per million input tokens and $6 per million output, the same as Grok 4.6.
What changed
xAI says it used a larger base model, a longer reinforcement learning run and training weighted toward hard tasks that can take hours. It also added SpaceX data, including Starlink telemetry, manufacturing records and engineering failure logs, aiming for a model that reasons better about hardware and physical systems.
Where it ranks
It went live without a waitlist on the xAI API, Cursor and GitHub Copilot. The release had been delayed five times since late July. On GDPval it scored 1695, second to Claude Fable 5.1 at 1735. Grok is aiming for strong results at a mid-range price rather than the top spot.