Intern-S2-Preview is a family of foundation models built for multimodal scientific reasoning and long-horizon agentic tasks. Its training pipeline combines scientific multimodal pre-training with supervised fine-tuning, scalable multi-task reinforcement learning, and both black-box and white-box agentic RL. The flagship 397-billion-parameter model achieves competitive results across scientific benchmarks, while a separate 4-billion-parameter memory-augmented extension improves performance on specialized biology tasks without altering the main model backbone.
