Vals raises $40 million to reshape AI model benchmarking
Vals, a startup founded in 2024, is developing a benchmarking system intended to make evaluations of artificial intelligence models more current and reliable. The company says many older academic tests no longer reflect the capabilities of newer models, while publicly available questions can allow developers to train specifically for those tests.
Rather than disclose its detailed test materials, Vals assesses models on complex, industry-specific work in areas including law, finance and coding. Its evaluations examine whether a model can produce work comparable in quality to a human’s, while also considering possible negative consequences. The company says it has developed assessments covering recursive self-improvement, mental health, cybersecurity, biosecurity and the law of armed conflict, including application of the Geneva Convention.
Companies pay Vals to evaluate their models, using the results to identify weaknesses and guide improvements. The startup recently raised $40 million in a Series A led by Andreessen Horowitz, following an earlier seed round led by 8VC and Bloomberg Beta. Vals said its revenue is eight times higher than last year’s level. Its staff has grown from eight people to 25, and the company plans to hire another 10 to 15 employees and expand its office. It has also launched a program for evaluating models for federal agencies.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.