QuixiAI releases QuixiMath-1B, a synthetic step-by-step math reasoning dataset on Hugging Face

QuixiAI · x · 2026-09-07

Quixi AI published QuixiMath-1B on Hugging Face, a synthetic dataset for step-by-step math reasoning (10M-100M rows, parquet, text-generation/language-modeling tasks). Each sample includes a problem (e.g., absolute-value and two-step inequalities), structured solution steps with typed labels, the final answer, plus metadata like generator type and grade-level difficulty — useful for training or evaluating stepwise math reasoning.

Related event: Eric Hartford Open-Sources QuixiMath Synthetic Math Reasoning Dataset(2 posts)→

Original post →

More from Research

Research channel →