Skip to content

g1

Build o1-like reasoning chains with a prompted open model

Visit official site
Category AI models

What g1 is

The project is described as an experimental effort to create o1-like reasoning chains with Llama-3.1 70b on Groq hardware, showing that prompting alone can improve logical problem solving.

What you can do with it

  • Visualise each step of the reasoning chain
  • Explore alternate answers through prompts
  • Apply more than one method to derive an answer

Getting started

  • Clone the project from its GitHub page
  • Read the README file for setup notes
  • Install dependencies and run the app.py demo

Pros & cons

✓ What we like

  • Several prompt strategies are supported
  • The reasoning steps are shown to the user

! What to watch out for

  • The project is described as experimental
  • It relies on Llama-3.1 70b running on Groq hardware

FAQ

Which model is used?

The material names Llama-3.1 70b on Groq hardware.

Is training needed?

The material states prompting is used without extra training.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review