Research

Is Grep All You Need? How Agent Harnesses Reshape Agentic Search

Frontier Coding Agents Can Now Implement an AlphaZero Self-Play Machine Learning Pipeline For Connect Four That Performs Comparably to an External Solver

Embarrassingly Simple Self-Distillation Improves Code Generation

Building AI Coding Agents for the Terminal: Scaffolding, Harness, Context Engineering, and Lessons Learned

Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models

Recursive Language Models

Pagina LLM Post-Training: A Deep Dive into Reasoning Large Language Models

Pagina SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion