Breaking

SpaceXAI Releases Grok 4.7 for Coding and Knowledge Work

September 21, 2026 at 12:21 EDT

Foundation Models
AI Agents
Software Dev & Coding

SpaceXAI introduced Grok 4.7 on September 21, 2026, calling it its most capable model for coding and knowledge work and keeping the standard API price at $2 per million input tokens and $6 per million output tokens. The release pushes the former xAI product line further into long-running agentic software tasks, with availability in Cursor, Grok Build, the Grok API, third-party coding harnesses, model routers and cloud platforms.

Sep 21, 2026 · SpaceXAI

Grok 4.7 moves into long-running work

The new flagship targets coding, knowledge work and agent workflows with a 500,000-token context window and standard API pricing of $2 input and $6 output per million tokens.

500k

token context window

$2 / $6

input and output per 1M tokens

62.4%

LatchBio biosafety score

3.3%

risky prompts allowed on HackerBench

Terminal-Bench 4.0: Grok closes one gap

Each stacked block represents roughly six percentage points; labels show the score.

20.3%

Grok 4.6

38.0%

Grok 4.7

37.3%

GPT-5.6 Sol

57.9%

Fable 5.1

What changed inside the model

Larger base
new model foundation

→

Longer RL
harder multi-hour tasks

→

Self-checking
better verification

→

Agent harness
built for Grok Bot

Why builders will try it

Low standard token prices, a large context window and agent-focused training make it practical to test on repositories and office work.

What remains unproven

Rival models still lead several rows, and general chat quality will need separate user testing.

Continue reading

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