INTELLECT-3: Technical Report
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915684325261312 |
|---|---|
| author | Prime Intellect Team Senghaas, Mika Obeid, Fares Jaghouar, Sami Brown, William Ong, Jack Min Auras, Daniel Sirovatka, Matej Straube, Jannik Baker, Andrew Müller, Sebastian Mattern, Justus Basra, Manveer Ismail, Aiman Scherm, Dominik Miller, Cooper Patel, Ameen Kirsten, Simon Sieg, Mario Reetz, Christian Erdem, Kemal Weisser, Vincent Hagemann, Johannes |
| author_facet | Prime Intellect Team Senghaas, Mika Obeid, Fares Jaghouar, Sami Brown, William Ong, Jack Min Auras, Daniel Sirovatka, Matej Straube, Jannik Baker, Andrew Müller, Sebastian Mattern, Justus Basra, Manveer Ismail, Aiman Scherm, Dominik Miller, Cooper Patel, Ameen Kirsten, Simon Sieg, Mario Reetz, Christian Erdem, Kemal Weisser, Vincent Hagemann, Johannes |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_16144 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | INTELLECT-3: Technical Report Prime Intellect Team Senghaas, Mika Obeid, Fares Jaghouar, Sami Brown, William Ong, Jack Min Auras, Daniel Sirovatka, Matej Straube, Jannik Baker, Andrew Müller, Sebastian Mattern, Justus Basra, Manveer Ismail, Aiman Scherm, Dominik Miller, Cooper Patel, Ameen Kirsten, Simon Sieg, Mario Reetz, Christian Erdem, Kemal Weisser, Vincent Hagemann, Johannes Machine Learning Artificial Intelligence 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. |
| title | INTELLECT-3: Technical Report |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2512.16144 |