Tether’s AI research division, operating under the name QVAC, has released Genesis III, a synthetic STEM dataset containing 191.43 billion tokens. A 1.7 billion-parameter model trained on a portion of that dataset hit a 99.45% valid-answer rate on MMLU STEM benchmarks, earning the accompanying research paper acceptance at COLM 2026. The dataset spans approximately 159.6 million documents across 19 different STEM domains, organized into three difficulty tiers: high school, college, and professional. The model trained on the Option-Level split alone produced the 99.45% valid-answer rate. Against the Cosmopedia-v2 baseline, Genesis III-trained models improved by 28.57 points on ARC-Easy, 21.35 points on ARC-Challenge, and 15.03 points on MMLU STEM.
Source: Read the original article
