About
The human behind the leaderboard.
I'm a dad of three kids living in the jungle, and I have absolutely no idea what's going on anymore.
Every week there's a new model, a new benchmark, a new company claiming they've solved AGI over the weekend. OpenAI drops something, then DeepSeek drops something bigger, then Google drops something sideways, then some lab I've never heard of in Shenzhen casually beats everyone on math.
I built The AI Race because I was drowning. 25 organizations. 24 benchmarks. 7 categories. 4 modalities. I needed one place to see who's actually winning — and by how much — without reading 47 blog posts and three Substacks.
This whole thing was vibe-coded between nap times and school runs. It's not perfect, but it works. And honestly, it tracks more AI benchmarks in one place than most things I've found — which is either impressive or sad, I'm not sure which.
If you find this useful, or if you just want to argue about rankings, come find me on X. Bug reports, hot takes, and unsolicited opinions about which model is actually the best are all welcome. Seriously — I'm a guy in the jungle, I'll take any conversation I can get.
Why This Exists
New frontier model every 48 hours. I can't keep up. You can't keep up. Nobody can keep up.
MMLU, GPQA, SWE-bench, Arena ELO, AIME, ARC-AGI... each one tells a different story.
US, China, EU — this isn't one company vs another. It's geopolitics with gradient descent.
Three kids, limited screen time, zero patience for scrolling through 12 leaderboard sites.
Built With
Vibes, caffeine, and the following:
Made with equal parts curiosity and confusion from somewhere in the jungle.