About

Independent AI Benchmarks

Tokonomix is an independent AI benchmark platform. No vendor relationships. No sponsored rankings. Pure data.

“The AI market moves faster than any single analyst can track. Tokonomix measures so you can decide.”

Founder & Editorial Lead

Mes Kalkan

Founder & Editorial Lead

Founder of InterIP Networks BV. Builds and maintains the Tokonomix benchmark methodology and signs off on every published methodology change.

Tokonomix grew out of a personal need to make evidence-based AI model selection decisions for InterIP customer projects. Every methodology change is signed, dated, and reversible.

mkalkan@interip.nlLinkedIn: TODO

How it works

01⚙️

We measure

Automated test runners call every active model API 4× per day from EU endpoints. Speed, quality, and availability are recorded per run.

02📊

We analyse

P50/P95 latency, quality scores (judge-LLM, blind), and error rates are computed and stored. No manual curation. No editorial spin.

03📡

We publish

Results are live within minutes of each test cycle. The full benchmark dataset is available as JSON for LLM crawlers and downstream tools.

Editorial standards

Model pages are compiled by the Tokonomix Editorial Team from automated benchmark runs and reviewed before publication. The methodology itself is owned and signed by Mes Kalkan. Material methodology changes are logged in a public changelog. Data corrections route through the methodology owner and are published within 24 hours of a verified report.

Read the methodology →Submit a correction →

Built by InterIP Networks BV

InterIP Networks B.V. is an independent infrastructure and technology company based in the Netherlands. We build tools for ourselves and release them publicly when they have broader value. Tokonomix was born from the need to make evidence-based AI model selection decisions — so we built the tool we wanted to use.

Legal entity: InterIP Networks BV
KvK (Chamber of Commerce): 27183782
Registered address: Linnewever 18, 2292 JH Wateringen, Netherlands
Founded: 2024

We are not affiliated with any AI provider. No investor relationships. No sponsored content.

Open data

All benchmark results are freely accessible. Use our data in your own research, tools, or dashboards — no key required.

Dataset & JSON export →Methodology →

Full dataset available as JSON at /api/md/{lang}/dataset.

Contact & corrections

Questions, corrections, or partnership inquiries:

mkalkan@interip.nl

Found a data error? Open an issue or send an email. We aim to correct benchmark data within 24 hours of a verified report.