According to CoinMeta, Vals has launched the AI evaluation platform Vals-Smith. This platform is capable of reading the GitHub code repository, extracting real development tasks from the merged PR, and generating customized programming evaluations. Each question includes a problem description, hidden tests, and a reproducible environment. The models being tested can only see the code before repairs and are not able to view the original patches. Teams can use this platform to have different models and coding agent run the same set of questions, allowing for a direct comparison of pass rates. Vals has also completed a $40 million Series A financing round, with a valuation of $400 million, led by a16z. The results from this platform have been referenced by models from OpenAI, Anthropic, Google, Meta, and xAI. This year's revenue has already reached 8 times that of the entire year of 2025.