Grok 4.7 Launches With Focus on Coding, Long-Running Tasks and Knowledge Work

21 Sep 2026

Published: 21 Sep 2026, 11:43 pm IST

xAI has officially launched Grok 4.7, its latest artificial intelligence model, with a stronger focus on coding, complex reasoning and professional knowledge work. The new model arrives as the successor to Grok 4.6 and is designed to spend more time working through difficult tasks rather than rushing to produce an answer.

According to xAI, Grok 4.7 uses a new and larger base model compared with Grok 4.6. The company also says it underwent a longer reinforcement-learning training process using a harder mix of tasks, particularly problems that can take several hours to complete.

Grok 4.7 Focuses on Complex Coding Tasks

One of the major areas of improvement is software development. xAI says Grok 4.7 has been trained to work more effectively on longer coding assignments and to check its own work more carefully.

The model is also designed to better manage longer contexts, which can be useful when developers are working on large codebases or multi-step projects. In CursorBench 4.0, xAI reports that Grok 4.7 achieved 46.3%, compared with 40.4% for Grok 4.6. It also reported a rise on Terminal-Bench 4.0 from 20.3% with Grok 4.6 to 38.0% with Grok 4.7. (Cursor – Community Forum)

These figures come from xAI’s launch material, so they are best viewed as company-reported benchmark results, rather than a universal independent ranking of AI models.

More Than Just a Coding Model

Although coding is a major focus, xAI is positioning Grok 4.7 as a broader knowledge-work model. The company says it has been trained to understand the Grok Bot harness natively, helping it handle conversational tasks and professional workflows.

xAI also says the model performs better when creating documents and presentations. This could make Grok 4.7 useful beyond software development, particularly for users who rely on AI for research, writing and other multi-step knowledge tasks.

Same Price and Speed as Grok 4.6

Another notable part of the launch is that xAI says Grok 4.7 is being offered at the same price and speed as Grok 4.6. The company lists pricing at $2 per million input tokens and $6 per million output tokens for the standard API pricing structure.

Grok 4.7 is available through the xAI API and has also been announced for platforms including Cursor and Grok Build.

For developers and AI users, the launch adds another major model option at a time when competition around coding agents, reasoning systems and long-running AI workflows is rapidly expanding. The more important test now will be how Grok 4.7 performs on independent evaluations and real-world workloads beyond the company’s launch benchmarks.

Also Read: Asian Games 2026 Live: India and Pakistan Schedule, Results, Medal Tally, Cricket, Hockey and Latest Updates