VibeThinker-1.5B by Sina Weibo
1.5B language model with strong reasoning at tiny cost
#Open Source
#Artificial Intelligence
#GitHub
SUMMARY
VibeThinker-1.5B by Sina Weibo – Efficient 1.5B-parameter language model with strong reasoning
Summary: VibeThinker-1.5B is a dense 1.5 billion parameter language model trained for $7,800 that delivers reasoning performance comparable to larger models like GPT OSS-20B Medium while remaining efficient and easy to run.
What it does
It provides strong reasoning capabilities through a compact, dense architecture trained at low cost, enabling efficient inference and deployment.
Who it's for
Researchers, startups, and edge deployment teams needing high reasoning performance with limited training and inference budgets.
Why it matters
It offers comparable reasoning to much larger models at a fraction of the training and operational cost.