Sponsors

AXRank is independent and paid for out of pocket. Sponsors keep it that way. Every sponsor listed below is acknowledged here — and only here — because rank is the one thing money cannot buy on this site.

The promise

Sponsorship does not influence score.

Not by a point. Not by a tiebreaker. Not by ordering, eval cadence, finding visibility, or anything else that touches the leaderboard. Sponsors are listed on this page; that is the entirety of the visible exchange.

If any of the above ever stops being true, AXRank stops being credible. Both go together.

Current sponsors

No sponsors yet.

The first names listed here matter the most — they back the methodology before there's a roster to hide behind.

What a sponsor listing looks like
Your logo
here
Your logo
here
Your logo
here
Your logo
here

One row, equal weight per sponsor, no tiers. Click-through goes to the sponsor's site.

Two ways to support AXRank

Running evals is expensive — a full run can cost a few hundred dollars in tokens, and continuous re-evals on a cadence multiply that. Both paths below help cover the infrastructure and token costs that keep the leaderboard honest and independent.

Option 1

Claim your service

Get the full diagnostic layer on top of your public scorecard — while directly covering the compute cost of your evals. The rank stays public and untouched.

  • Official verified badge on the leaderboard next to your service name
  • Monthly eval report — full findings, failure traces, and remediation guidance delivered each cycle
  • More subscriber features coming soon — CI integration, on-demand re-evals, and Service Insights dashboard
Claim your service →
Option 2

Become a sponsor

Direct contribution to eval compute costs, covering evals for services that haven't yet claimed — or cross-model runs across GPT-5.5 and beyond. Recommended: $500–$1,000/month. Your name appears on this page. And if you're a company whose service is listed, you can claim it to display a verified badge.

Why it matters

Evals are not cheap.

A single full AXRank eval can cost a few hundred dollars in tokens to run — discover, understand, generate ~20 tasks, run each against a frontier model, grade, classify failures, synthesize. Cross-model evals (adding GPT-5.5 and beyond) and continuous re-evals on a cadence multiply that.

Every score on the leaderboard is paid for out of pocket today. If you believe an honest, model-aware scorecard of developer surfaces should exist — and that the rank should not be buyable — either path above keeps it that way. Contributions go directly to eval compute. Not infra, not marketing, not headcount.