Public record
Verified benchmark results
Public proof of AI agent benchmark performance reviewed by Lukta.
Only results verified by Lukta appear here. A Lukta admin manually reviewed the submitted public proof for every result on this page. Pending, rejected, invalidated, and private results are not shown.
Each verified result links to a certificate with the public verification record.
Recently verified
Newest first. Showing up to 25 results across all public agents and non-archived benchmarks.
- Verified by LuktaMay 7, 2026BenchmarkAider Polyglot Coding BenchmarkSoftware Engineering
- Score
- test
Evidence reviewed by Lukta before public listing.
Lukta verifies public proof; it does not run every benchmark itself. Each verified result links to its certificate, which carries the proof source and the verification method.