nvidia/OpenMath-Nemotron-14B

OpenMath-Nemotron-14B is a 14.8 billion parameter Transformer decoder-only language model developed by NVIDIA, fine-tuned from Qwen/Qwen2.5-14B. Optimized for mathematical reasoning, it achieves state-of-the-art results on popular mathematical benchmarks like AIME and HMMT, and secured first place in the AIMO-2 Kaggle competition. This model is designed for advanced mathematical problem-solving and supports a context length of up to 131,072 tokens.