🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model! ✨ Built on V3.1-Terminus, it debuts DeepSeek Sparse Attention(DSA) for faster, more efficient training & inference on long context. 👉 Now live on App, Web, and API. 💰 API prices cut by 50%+! 1/n

Sep 29, 2025, 10:10 AM UTC

Views: 1,424,571
Likes: 6,919
Replies: 321
Reposts: 871
Quotes: 354
Bookmarks: 1,055

Xtext

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model!

DeepSeek announces an experimental model built on V3.1-Terminus. The post highlights a new sparse-attention feature, availability across App, Web, and API, and a price reduction of more than 50% for API use.

Views: 1.4M
Likes: 6.9K
Replies: 321
Reposts: 871
Quotes: 354
Bookmarks: 1.1K
Topics: Experimental model launchSparse attentionAPI pricingLong-context processingAI

Analysis

Content overview

Formats

Product AnnouncementFeature LaunchThread Opener

Topics

Experimental model launchSparse attentionAPI pricingLong-context processing

Sign in to see the full analysis.

Full post text

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model! ✨ Built on V3.1-Terminus, it debuts DeepSeek Sparse Attention(DSA) for faster, more efficient training & inference on long context. 👉 Now live on App, Web, and API. 💰 API prices cut by 50%+! 1/n

Tweet text hook

As written

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model!

Tweet text structureThe Product Launch Drop

From the post

01
The Announcement
Introduce the new offer with energy and frame it as a fresh or experimental release.
02
The Main Benefit
Explain what the offer adds and connect it to a clear improvement in performance or efficiency.
03
Availability and Value
Tell people where they can use it and highlight a concrete price or access advantage.
04
The Continuation Marker
Signal that the announcement is part of a longer thread or sequence.