Launching today
PAEval
Compare personal AI agents with source-linked evidence
1 follower
Compare personal AI agents with source-linked evidence
1 follower
Compare personal AI agents using published benchmarks and documented capabilities. PAEval tracks 88 products and keeps each evaluator’s original scales, samples and configurations visible. Explore side-by-side comparisons, capability evidence and source links. Missing evidence stays explicit. The PAEval Index is in development; no composite ratings are published.

Free
Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
I maintain PAEval, a comparison site for personal AI agents. It brings existing third-party evaluations and product documentation together, with links back to the original evidence.
The November 22 snapshot tracks 88 products. Fifty-five have source-backed capability profiles; 33 have external scores compiled and are still awaiting those profiles.
The comparison tool keeps task results and capability evidence alongside one another. You can follow a score to its publisher and see its original metric, sample and configuration. Unknown, conditional and unsupported capability states are kept distinct.
PAEval does not run new evaluations. The evaluators’ results are not directly interchangeable, and the composite PAEval Index is still in development. No Index ratings are published.
Try a comparison at https://paeval.com/compare or browse the source directory at https://paeval.com/benchmarks.
I would particularly appreciate corrections to source attribution, missing product configurations and feedback on which comparison questions the site does not answer clearly.