Breaking

Elon Musk Says Latest Grok Progress Makes AGI Feel Closer as SpaceXAI Pushes Agents

September 27, 2026 at 22:01 EDT

Foundation Models
AI Agents
Software Dev & Coding

Elon Musk has again put xAI's Grok roadmap in the AGI spotlight, saying the latest progress made artificial general intelligence feel closer than before. The remark is not a product launch or proof that AGI has been achieved, but it arrives as SpaceXAI is moving Grok from benchmark-focused models toward longer-running agents.

Sep 28, 2026 · SpaceXAI

Grok 5 Returns to the AGI Debate

Musk's latest AGI remark lands after Grok 4.7 and Grok Bot pushed the roadmap toward longer-running agents, while Grok 5 remains framed around a 6 trillion-parameter target.

The stated Grok 5 scale doubles the earlier base

Each block represents 1 trillion parameters.

3TGrok 3 and 4 base

6TGrok 5 target

10%

Musk's stated AGI chance estimate for Grok 5

71.0%

Grok 4.7 score on DeepSWE v1.1

$2/$6

Starting input/output price per 1M tokens

From model gains to agent work

Grok 4.7
coding and knowledge work

→

Grok Bot
persistent tool use

→

Grok 5
multimodal AGI test

Why developers are interested

Grok 4.7 is praised for codebase navigation, debugging, and favorable price-performance when tasks are clearly scoped.

Why the claim is disputed

Critics cite verbosity, unwanted code changes, planning gaps, and higher cost per finished task in some workflows.

Continue reading

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