Sail model boosts scientific research with fewer parameters
SAIL: Scientific Agentic Intelligence via a Science-Aware Loop
Computation and LanguageDigital LibrariesInformation Retrieval
Summary
Finding and using scientific papers and code can be very tricky and slow. The authors created SAIL, an AI system that improves itself by learning from its own mistakes when handling literature review, coding, and research planning tasks. It studies its failures, creates new training problems to fix them, and repeats this process to get better. This approach lets SAIL work well on complex science tasks while being smaller and more efficient than similar models.
What this means in practice
- •For scientific software developers: Build smarter research tools that understand and create scientific code with fewer computational resources.
- •For technical content analysts: Automate literature review processes by evaluating and selecting relevant scientific evidence more effectively.
Authors
SAIL Model Team, Boyuan Sun, Bryan Dai, Che Liu, Chi Liu, Derek Li, Hongming Piao, Mengzhuo Chen, Xidong Wang, Yan Shu, Yinda Chen, Ziyang Zeng
Abstract
We introduce SAIL, an open model with 35B total and 3B active parameters for literature research, scientific coding, and multi-step research workflows. SAIL is developed through a science-aware improvement loop: agents built on frontier AI models analyze its task failures and construct training tasks that address the underlying capability gaps. The diagnosis examines search and evidence selection in literature tasks, scientific assumptions and reasoning in coding, and planning and revision in longer investigations. The agents draw on paper collections and scientific code repositories to build problems, interaction trajectories, and executable tasks with the required environments and tools. We repeat this loop over multiple development cycles and train SAIL through supervised fine-tuning, specialist training, multi-teacher on-policy distillation, and agentic reinforcement learning. SAIL achieves competitive performance across scientific research tasks with substantially fewer parameters than leading open-weight models.