No token math, no per-model price sheets. Each message costs exactly one request from your window — from the fastest chat model to the deepest Alloy reasoning. Free tier included, no card ever.
https://api.atlasbrain.cloud/v1
–models live
1 requestflat cost per message
100%verified accuracy, top Alloy tiers*
$0to start
The Alloy family
Our own frontier models, built and refined in-house: they reason deeply, self-review, and verify before answering — and report how sure they are. Accuracy figures are from our published exact-match benchmark, re-run and posted openly.
Frontier catalog
Direct access to the latest frontier chat models through the same endpoint and key. This list is live — it is exactly what GET /v1/models returns right now.
Loading live catalog…
What you get without paying
Rolling windows, not monthly caps — and you can top the free tier up any day by watching an ad, finishing a quick link, or claiming the daily bonus.
Plan
5-hour window
7-day window
Requests / min
Concurrent
Free
50
200
20
2
Pro
500
3,000
100
8
Premium
2,500
10,000
500
24
*Accuracy: 100% (25/25) on our exact-match benchmark for Alloy at low & medium thinking and Alloy Pro at high thinking (96% for code sessions and Alloy Pro at medium) — method, latency, and honest caveats on the benchmarks page. Alloy and the full frontier catalog work on the free tier; Alloy Pro requires Pro or Premium.