Support the benchmark

Support ARI Bench

Help fund independent benchmark infrastructure for the capability that matters after human parity: AI improving AI.

What supporters make possible

A benchmark this important needs independence, repeatability, and enough operational depth to stay credible as models improve.

Funding helps maintain hidden evaluation surfaces, secure handling, and independent benchmark administration.

Support enables model sweeps, validation runs, replication work, and faster publication of credible results.

ARI can publish clearer reports, better methodology notes, and stronger validation without depending on one lab.

Individual or company support for benchmark maintenance, public reporting, and future evaluation tracks.

Fund public prizes or challenges for systems that demonstrate meaningful recursive improvement.

Support methodology, analysis, validation, safety review, and interpretation of recursive AI improvement signals.

Contribute compute, orchestration, storage, monitoring, or security infrastructure for benchmark operation.

Support public reports that explain results, leaderboard movement, and what progress means for the field.

Discuss larger funding, partnership, or long-term benchmark development support with ARI Bench directly.

Some supporters may prefer to fund benchmark infrastructure without public recognition. Private support can be discussed directly.

Support contributions may not be tax deductible unless a fiscal sponsor arrangement is explicitly confirmed.

If you want to support ARI Bench, send a short note with the type of support you have in mind and whether you want public recognition.

Founding supporters Prize funding Infrastructure support