Olmo 3.1 32B Think
Allen Institute for AI (Ai2) · 2025-12-12 · 32.2B parameters
Olmo 3.1 32B Think is Ai2's fully open 32B reasoning model, produced by continuing the Olmo 3 Think reinforcement-learning run for an additional 21 days on 224 GPUs. It generates long chain-of-thought reasoning for math, coding, and other multi-step tasks, and Ai2 releases its weights, training data, code, and intermediate checkpoints.