Niklas Muennighoff
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Inside s1: An o1-Style Reasoning Model That Cost Under $50 to Train with Niklas Muennighoff
- Published
- March 3, 2025
- Duration
- 49:29
- Summary source
- description
- Last updated
- Jun 7, 2026
Discusses openai, google-ai.
Summary
Today, we're joined by Niklas Muennighoff, a PhD student at Stanford University, to discuss his paper, “S1: Simple Test-Time Scaling.” We explore the motivations behind S1, as well as how it compares to OpenAI's O1 and DeepSeek's R1 models. We dig into the different approaches to test-time scaling, including parallel and sequential scaling, as well as S1’…
Intelligent report
Sign in to read teasers, or upgrade to Research Pro to commission a new dossier for this episode. Learn more →
Show notes
Today, we're joined by Niklas Muennighoff, a PhD student at Stanford University, to discuss his paper, “S1: Simple Test-Time Scaling.” We explore the motivations behind S1, as well as how it compares to OpenAI's O1 and DeepSeek's R1 models. We dig into the different approaches to test-time scaling, including parallel and sequential scaling, as well as S1’s data curation process, its training recipe, and its use of model distillation from Google Gemini and DeepSeek R1. We explore the novel "budge
Themes
- openai
- google-ai