INTELLECT-3: A 100B+ MoE trained with large-scale RL

Today, we release INTELLECT-3, a 100B+ parameter Mixture-of-Experts model trained on our RL stack, achieving state-of-the-art performance for its size across math, code, science and reasoning benchmarks, outperforming many larger frontier models.

Our complete recipe — from the model weights and training frameworks, to our datasets, RL environments, and evaluations — has been open-sourced, with the goal of encouraging more open research on large scale reinforcement learning.