Get Started
Home
Topics
Search
Library
Research questionHow can language models solve difficult algorithmic programming problems under realistic contest time, access, and submission constraints?Contest programming requires both algorithmic reasoning and executable code that passes the problem’s tests. Language models may need specialized training and additional inference computation to meet these demands within strict contest limits.
Code Generation & Program Synthesis
Evaluation & Benchmarks
Inference Optimization
LLM Pretraining & Post-training
Reasoning
Reinforcement Learning
Latest papersRecent research connected to this question, newest first.Post-Training Language Models for Gold-Medal Performance in Coding CompetitionsThe paper combines curated problem data, synthetic reasoning traces, supervised and reinforcement-learning post-training, and iterative test-time generation and refinement. It evaluates specialized systems on IOI-style problem sets, including a prospective contest evaluation under human-contestant constraints; evidence is limited to those evaluations.research paper · Sep 4, 2026
Related questions
How can language models reliably follow instructions containing many simultaneous constraints?How can large language models allocate reasoning computation to preserve accuracy under limited training and inference budgets?How can large language models cut training and inference costs without materially harming accuracy?How can LLMs synthesize executable machines and improve their designs through physical simulation feedback?