🏅 CAIA Benchmark Leaderboard

This leaderboard shows the performance of various language models on the CAIA benchmark. The benchmark evaluates models both with and without access to tools.

Performance without tools: accuracy, Pass@k, cost, and cost efficiency across all models.

Performance without tools: accuracy, Pass@k, cost, and cost efficiency across all models.

Grok 4 Fast Reasoning	without_tools	0.118	12.9	16.9	0.0334	0.0593