CodeForces

The Codeforces rating evaluates models on competitive programming problems from Codeforces contests. The output is an Elo rating comparable to the human ladder — for scale: a professional competitive programmer sits far above 2000, world-class participants above 3000.

CodeForces measures hard reasoning in science, technology, engineering and math. In detail, it checks:

• algorithmic problems under real contest conditions
• efficient solutions that pass all hidden test cases within limits
• implementation skill: complete programs, not fragments
• head-to-head comparable Elo so ratings across models match

The score is an Elo rating, built from head-to-head comparisons between models (like chess ratings). The number is only meaningful relative to the other rated models — higher is better.

What is the CodeForces benchmark?

The Codeforces rating evaluates models on competitive programming problems from Codeforces contests. The output is an Elo rating comparable to the human ladder — for scale: a professional competitive programmer sits far above 2000, world-class participants above 3000.

What does CodeForces test?

CodeForces mainly measures hard reasoning in science, technology, engineering and math. Concretely, it checks: algorithmic problems under real contest conditions; efficient solutions that pass all hidden test cases within limits; implementation skill: complete programs, not fragments.

How is the CodeForces score calculated?

The score is an Elo rating, built from head-to-head comparisons between models (like chess ratings). The number is only meaningful relative to the other rated models — higher is better.

Where do I get CodeForces data and scores?

The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/codeforces/scores) or pull it directly from the Open LLM Radar API with any HTTP client.