BenchLM

BenchLM provides a comprehensive leaderboard for large language models, tracking 284 LLMs across 320 benchmarks. The platform offers detailed comparisons, including data on model quality, cost, and runtime, to assist in evaluating performance. The leaderboard features both supported and estimated models, presenting verified and provisional rankings based on the BenchAlign method. It includes real pricing and runtime information, enabling users to assess tradeoffs between various frontier AI models.