INTELLECT-3: Technical Report
This work addresses the problem of scalable and efficient reinforcement learning infrastructure for AI researchers and developers, with incremental improvements in model performance and open-source tooling.
The researchers tackled the challenge of creating a large-scale reinforcement learning infrastructure and model, resulting in INTELLECT-3, a 106B-parameter Mixture-of-Experts model that achieves state-of-the-art performance for its size across math, code, science, and reasoning benchmarks, outperforming many larger frontier models.
We present INTELLECT-3, a 106B-parameter Mixture-of-Experts model (12B active) trained with large-scale reinforcement learning on our end-to-end RL infrastructure stack. INTELLECT-3 achieves state of the art performance for its size across math, code, science and reasoning benchmarks, outperforming many larger frontier models. We open-source the model together with the full infrastructure stack used to create it, including RL frameworks, complete recipe, and a wide collection of environments, built with the verifiers library, for training and evaluation from our Environments Hub community platform. Built for this effort, we introduce prime-rl, an open framework for large-scale asynchronous reinforcement learning, which scales seamlessly from a single node to thousands of GPUs, and is tailored for agentic RL with first-class support for multi-turn interactions and tool use. Using this stack, we run both SFT and RL training on top of the GLM-4.5-Air-Base model, scaling RL training up to 512 H200s with high training efficiency.