ADGAI API
One API key, unlimited calls*, for use in your own apps or in the Ralph build loop — instead of metering every token against a per-request bill.
R1,000
per month
- One key across chat, code and structured-output endpoints
- Usable inside the Ralph loop for autonomous build/test/fix cycles
- Usable in your own product's backend
- Flat monthly rate — no per-token invoice surprises
- Cancel anytime, month to month
* Unlimited within a fair-use ceiling for a single application. High-volume or multi-tenant use is scoped and quoted separately — talk to us before you integrate at scale.
FOR CONTEXT
What the underlying models cost per token, today.
A flat monthly key only makes sense if you understand what metered pricing looks like underneath it. Here's what a few current models cost directly from their own providers.
| MODEL | PROVIDER | INPUT | OUTPUT |
|---|---|---|---|
| Claude Haiku 4.5 | Anthropic | $1 / MTok | $5 / MTok |
| Claude Sonnet 5 | Anthropic | $2 / MTok | $10 / MTok |
| Claude Opus 5 | Anthropic | $5 / MTok | $25 / MTok |
| Qwen3.7 Flash | Alibaba Cloud | $0.03 / MTok | $0.13 / MTok |
| Qwen3 Max | Alibaba Cloud | $0.78 / MTok | $3.90 / MTok |
| Qwen3.8 Max | Alibaba Cloud | $2.00 / MTok | $6.00 / MTok |
MTok = million tokens, list prices in USD from each provider's own published pricing page at time of writing. Prices change — check the provider directly before budgeting against these. ADGSTUDIOS is not affiliated with Anthropic or Alibaba Cloud; this table is for orientation, not a resale offer of their APIs.
Source: Anthropic — Claude API pricing. Qwen pricing per Alibaba Cloud Model Studio's published international rates.
ON BENCHMARKS
We're not going to publish our own intelligence benchmark numbers here without running them ourselves. For independent, regularly-updated model benchmarks, see third-party trackers such as Artificial Analysis or the LMArena leaderboard.