Chatbot Arena ELO rating for math-related prompts — human-judged mathematical reasoning quality.
Adds human-preference signal for explanation clarity and usefulness in math responses; complements static correctness benchmarks with live prompt diversity.
Many voters cannot verify advanced derivations, so polished style can outperform correctness; prompt mix is uncontrolled and may overweight simpler math tasks. Data: Arena (lmarena.ai) leaderboard dataset, https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset, licensed CC-BY-4.0; rows are converted to the ELO fields shown here.
Higher is better. ELO-like rating for math tasks.
Many voters cannot verify advanced derivations, so polished style can outperform correctness.
No scores available for this benchmark yet.