New Benchmark Exposes How Badly Vision-Language Models Describe Charts — and Proposes a Rubric That Actually Measures What Matters

New Benchmark Exposes How Badly Vision-Language Models Describe Charts — and Proposes a Rubric That Actually Measures What Matters | BedrockNews