Vals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking

Val's, an artificial intelligence benchmarking startup founded in 2024, is aiming to be the benchmarking standard in the industry. Led by co-founder Rayan Krishnan, the company recently secured a $40 million Series A round of funding led by Andreessen Horowitz. Prior to this, Krishnan interned at Palantir and worked with Microsoft and Stanford's AI lab during his undergraduate studies.

Benchmarking plays a crucial role in validating the capabilities of AI models, which can significantly impact a company's reputation. However, many existing benchmarking systems are outdated and struggle to measure the performance of modern models. Vals seeks to address this by focusing on real-world applications, rather than abstract intelligence tests.

The company's tests are designed to assess an AI model's ability to perform complex tasks in industries such as law, finance, coding, mental health, cybersecurity, biosecurity, and even law of armed conflict.

Unlike many benchmarking systems that publicly disclose test materials, allowing companies to train their models against them, Vals keeps its test materials confidential. Instead, the startup evaluates models based on their ability to complete tasks that produce results comparable to human performance, both positive and negative. Krishnan believes that understanding the potential downsides of AI models is just as important as identifying their positive impacts.

As Vals continues to grow, it is increasingly becoming a key decision-making factor for companies evaluating AI models. Krishnan compares the revenue model to students paying the College Board for SAT testing, emphasizing that effective evaluation helps companies improve their AI models over time. The startup's revenue has grown eightfold over the past year, and its team has quadrupled to 25 employees.

Krishnan envisions Vals as the future of AI benchmarking, as AI models become more integrated into society and play a central role in the economy. With companies like SpaceX, Anthropic, and OpenAI preparing for IPOs, the demand for robust benchmarking systems is expected to rise.