GOOGLE OPEN SOURCES LMEVAL FOR CROSS-MODEL EVALUATION

Summary from the TLDR AI newsletter

LMEval is a framework for benchmarking AI models across different providers with multimodal support and incremental evaluation.