Vals AI AI features and pricing (October 2026)

Vals AI builds benchmarks that test whether AI models can do professional work, such as legal research, financial modeling, tax work and medical coding, and publishes the results by domain.

Benchmarks are grouped by domain: legal (Legal Research Bench, Harvey's Legal Agent Benchmark, LegalBench), finance (Finance Agent v2, Excel Modeling Benchmark, Tax Agent Bench), healthcare (MedCode, MedScribe), math, science, education, voice, coding and more. Many are proprietary, others come from industry partners or academia. According to its about page, Vals has built and open-sourced its evaluation infrastructure, including Valkyrie, a distributed system for running agentic benchmarks, and a model library that offers a free standard API for calling models.

• Tests models on finance tasksFinance Agent v2, the Excel Modeling Benchmark and Tax Agent Bench show how models handle analyst, modeling and tax work. Finance leaders can see where models are usable and where people must stay in charge.
• Compares models for legal researchLegal Research Bench and LegalBench results help legal teams judge models for research and drafting support before trusting them with client work. Results are shown per benchmark, so teams can look at the tasks closest to their own practice.
• Weighs accuracy against costResults include latency, cost and standard errors, so teams can see whether a more expensive model is meaningfully better on a given professional task. This helps keep running costs in line with the value of the task.

• Benchmarks organized by professional domain, including law, finance, tax and healthcare.
• Benchmarking infrastructure Valkyrie released as open source.

Vals AI lists an office in San Francisco on its linked careers page; its legal and tax benchmarks focus on US professional work.

Legal and tax benchmarks reflect mostly US practice; results do not show how models handle German law or German tax rules. Private test sets cannot be inspected. If you work with Vals on evaluations, clarify terms, data handling and whether your own documents are processed.

No reviews yet. Be the first to share how your team uses it.

Every review shows its verification level. We read each one before it goes live. How we check reviews