1 paper
Yueqi Song, Lintang Sutawika, Jiarui Liu +8
Evaluating LLM agents on benchmarks like SWE-Bench and GAIA can be expensive, time-consuming, and requires complex infrastructure. A single evaluation can cost thousands of dollars…