varies

The 'varies' tool appears to function as an evaluation dashboard comparing the performance of different large language models (LLMs) on the SWE-bench benchmarks for reasoning and software engineering tasks. It provides detailed statistics such as percentage resolved, cost per query, and allows users to explore results for various agents/models across multiple curated evaluations (full suite, multilingual, lite, multimodal). It is designed primarily for researchers, developers, and organizations benchmarking language models for code and reasoning tasks.