AI Disruption

AI Disruption

Claude Sonnet 4.5: 30+ Hours, 11K Code Lines

Claude Sonnet 4.5: 30+ hour coding, 11K lines, SOTA performance. VS Code extension & AI Agent SDK launched.

Meng Li's avatar
Meng Li
Sep 30, 2025
∙ Paid
3
2
Share

“AI Disruption” Publication 7800 Subscriptions 20% Discount Offer Link.


Claude Sonnet 4.5 Is Finally Here | by Jim Clyde Monge | Sep, 2025 |  Generative AI

Models are starting to compete intensively again!

Yesterday, DeepSeek open-sourced their new model V3.2-Exp, and Anthropic wasn’t far behind, launching the heavyweight Claude Sonnet 4.5.

Image

As the king of programming domains, Claude’s new model remains powerful, claiming to be the world’s best coding model.

We all know that GPT-5 Codex once claimed it could run independently for over 7 hours. But this time, Claude Sonnet 4.5 has pushed autonomous coding duration to over 30 hours.

Additionally, Claude claims it’s the strongest model for building complex agents, the best model for computer use, and shows tremendous progress in reasoning and mathematics.

Claude Sonnet 4.5 makes all of this possible. Anthropic released it alongside a series of major product upgrades:

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2025 Meng Li
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture