Compare models and prices yourself
Vals AI Model Benchmarks
Made by Vals AI
Free independent evaluations on real professional tasks — the place to check anyone's claims.
What it is
Vals AI runs its own evaluations of hosted models on actual professional work — legal reasoning, tax, finance, medicine — and publishes the results in full, without a paywall or an account.
This is the honest counterweight to every page like this one, including mine. When a company says its model is best at something, this is where you look to see whether an outsider measured the same thing.
Everything published is free to read. The company's paid work is private evaluation for businesses, with no figure published.
Published price
- Benchmarks and reportsNo paywall, no account needed$0
- Private evaluations for companiesA consulting engagement, so there is no list price to publishQuoted on request
Source: the Vals AI benchmarks · Last checked 17 September 2026
Benchmarks
The published figures for Vals AI Model Benchmarks, copied exactly as their sources give them and checked 17 September 2026. Each one names the test, who ran it and when, and says when the number comes from the maker rather than an independent tester.
No published score. Vals AI runs the tests, so it has no score of its own. As an example of what it publishes: the top model on its LegalBench board scored 88.56% when I checked.
Where to read today's numbers
Scores move, so these are the boards that keep them current.
- Vals AI benchmarksIndependent
Independent evaluations on real professional tasks — law, finance, tax, medicine — rather than vendor demos.
- Artificial AnalysisIndependent
Independent testing of quality, speed and price per million tokens across hosted models.
- LMArena leaderboardIndependent
Ranks models by blind head-to-head votes from the public. Good for a feel for general quality, weak on specialist work.
How it performs, in my experience
Independent evaluations on real professional tasks — law, finance, medicine — rather than vendor demos.
Free value
5 / 5
How much you get without paying.
Performance
4 / 5
How well it does its main job.
Range of uses
2 / 5
How many different jobs it suits.
Free independent evaluations — the place to check anyone's claims, mine included. These scores are my own judgment, not a measurement — the benchmark links above are the independent version.
Prompting it from this site
Can't be run from here
A benchmark publisher, not a model.
Use it for
- Checking a vendor's performance claim
- Picking a model for a specific kind of work
Running it yourself
Hosted only — nothing to install
Vals publishes benchmark results. It is a set of leaderboards to read, not software to install.
What you need first
- A browser
Step by step
1.Read the leaderboards on their site — nothing to install