AI Disruption

AI Disruption

Share this post

AI Disruption
AI Disruption
DeepSeek V3.1: Beats Claude 4 at Coding, But Not Writing

DeepSeek V3.1: Beats Claude 4 at Coding, But Not Writing

DeepSeek V3.1: 685B open-source AI beats Claude 4 in coding, excels in logic & long-context tasks, but writing lags behind.

Meng Li's avatar
Meng Li
Aug 20, 2025
∙ Paid
13

Share this post

AI Disruption
AI Disruption
DeepSeek V3.1: Beats Claude 4 at Coding, But Not Writing
5
Share

"AI Disruption" Publication 7400 Subscriptions 20% Discount Offer Link.


Deepseek Teases a Comeback!

Since the release of GPT-5, DeepSeek founder Liang Wenfeng has become the "busiest" person in the AI circle.

Netizens and media outlets keep pushing for updates every few days, either putting "pressure on Liang Wenfeng" or saying "the entire internet is waiting for Liang Wenfeng to make his move." Although we didn't get to see R2, DeepSeek officially launched and open-sourced its new model DeepSeek-V3.1-Base today.

Compared to Altman still painting grand visions of GPT-6 in today's interview, DeepSeek's new model arrival seems quite zen-like, with even the version number looking like a "minor patch."

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

Share