Breaking

SpaceXAI Launches Grok 4.7 for Coding Agents, Matching Grok 4.6 Pricing

September 21, 2026 at 18:16 EDT

Foundation Models
AI Agents
Software Dev & Coding

SpaceXAI released Grok 4.7 on September 21, presenting it as its most capable model for coding, agentic tasks and knowledge work while keeping the base API price at $2 per million input tokens and $6 per million output tokens. The model is available through the Grok API, Cursor, Grok Build, third-party coding harnesses, model routers and cloud platforms, according to the launch announcement.

Sep 21, 2026 · SpaceXAI

Grok 4.7 Moves Into Agent Work

The new model raises long-running coding and knowledge-work scores while keeping Grok 4.6 base pricing. The strongest gains appear in agent workflows, with cost and token use still central to adoption.

38.0%

Terminal-Bench 4.0

71.0%

DeepSWE v1.1

500k

Context window

$2/$6

Input/output per 1M tokens

Terminal-Bench nearly doubles

20.3%

Grok 4.6 High

38.0%

Grok 4.7 xHigh

A 17.7 percentage-point jump on the same benchmark family.

What changed under the hood

Larger base model→Longer RL on multi-hour tasks→Stronger self-verification→New safeguard stack

Positive read

Competitive long-task coding scores, stronger communication for design work, and unchanged base pricing.

Caution read

Higher effort can consume far more output tokens, and some benchmark comparisons mix effort levels.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.