Summary:
- Val, an AI benchmarking startup, has secured funding from Andreessen Horowitz to develop standardized metrics for evaluating large language models.
- The company aims to address the industry-wide challenge of "benchmark contamination" by providing reliable, rigorous testing environments to help enterprises measure AI performance and safety.