Hacker News 5 hours ago

VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO

Summary

Score: 137 points

Read Full Article

This is a preview of the article. Click below to read the full article on Hacker News.

Visit Hacker News

Article Details

Source: Hacker News
Category: Technology
Published: Tuesday, June 23, 2026
Time: 02:01 AM

Share this article

Twitter LinkedIn

More from Hacker News

Hacker News 1 hour ago

The new HTTP QUERY method explained

Hacker News 2 hours ago

Show HN: A pure ARM64 Assembly web server, now on Linux with CGI for no reason

Hacker News 3 hours ago

Will It Mythos?

Hacker News 4 hours ago

Ultralytics YOLO26: Unified Real-Time End-to-End Vision Models