Every frontier model.
One API, one invoice.
Claude, GPT, Gemini, DeepSeek and the video labs at wholesale rates. Billed in USD or AUD, with Australian regions when you need them.
Anthropic
OpenAI
Google
DeepSeek
Moonshot AI
Alibaba Qwen
ByteDance Seedance
Kling
xAI
Mistral
| Model | Region | Input | Output |
|---|
The rate card is the product
Provider list prices for reference, your rate below them, in USD or AUD. Every model and region on one key, one invoice.
See all models →Routes that fail over
Each model has more than one route. If one degrades, your request moves to the next, inside the regions you allow.
How routing works →Data policies you control
Pin regions, set retention, exclude any provider. Prompts only go where you allow, and every model is labelled with where it runs.
Data residency →Two lines to switch.
Point your existing OpenAI SDK at Seedrise and keep the rest of your code. Model IDs follow the providers' own names.
from openai import OpenAI
client = OpenAI(
- base_url="https://api.openai.com/v1",
+ base_url="https://api.seedrise.ai/v1",
+ api_key=os.environ["SEEDRISE_API_KEY"],
)
r = client.chat.completions.create(
model="anthropic/claude-opus-5.5",
messages=[{"role": "user", "content": "Summarise this contract."}],
)
# same call, any model on the rate cardModels and reference rates
Provider list prices in USD per 1M tokens, for reference. Your Seedrise rate card comes in below these.
Generative media, on the same key
Video and image generation from the labs shipping fastest. Same API, same invoice as your text models. Per-second and per-image list prices, your rate below them.
All image and video rates →Inference you can keep in Australia.
Models marked AU can be pinned to Sydney or Melbourne. Everything else is labelled with where it runs, so your data map stays honest.
- Claude Opus, Sonnet and Haiku 5.5Sydney · Melbourne
- Gemini 3.5 FlashSydney
- DeepSeek V3.2, Kimi K2.5, Qwen3 235BSydney
- Prompts and outputs never used for trainingEvery model
- No content retention by Seedrise, encrypted in transit and at restDefault
- Zero-retention routing, usage logs for audit, data processing agreementOn request
How pricing works
Every plan is priced below the providers' list prices. The more you run, the further below list you sit. Rates are quoted per account, in writing.
- No monthly minimum, no setup fee
- Monthly invoice in USD or AUD, or prepaid credit
- All models, all regions on the rate card
- Email support, one business day
- Everything in Pay as you go
- Usage and spend by project on the invoice
- Higher rate limits across all models
- Priority support with a named contact
- Everything in Volume
- Reserved capacity in Australian regions
- Data processing agreement and custom policies
- Private endpoints and migration help
Enterprise
A contract, a data processing agreement and a person to call. We set up the account, migrate your prompts and stay on the line.
Private endpoints
Dedicated base URLs and keys per business unit, with their own region and retention policy.
Custom volume pricing
Committed-use rates on the models you actually run, in AUD or USD, with usage by cost centre.
Priority support
A named engineer, a shared channel and a response time in writing. Business hours in AEST.
Migration assistance
We port your prompts, evals and model mappings from your current provider and run them side by side.
Questions buyers ask
Short answers. The rate card and terms spell out the rest.
Something not covered? [email protected]
How is my rate set? +
Can I choose where a model runs? +
Is there a minimum commitment? +
How does billing work? +
Is the API OpenAI-compatible? +
Can I trial it before signing anything? +
What support do you offer? +
What about the Chinese models? +
Get a rate card
Tell us what you run and we will send per-model rates, regions and terms.
- 01Tell us your stackWhich models, rough monthly volume, and whether you need Australian regions.
- 02Receive your rate cardWithin one business day. Per-model rates, regions and billing terms, in writing.
- 03Go liveA trial key first, then swap the base URL and keep your SDK. Invoices by project from day one.