Pricing
You pay the provider rate. The invoice never surprises you
Tokens bill per million at the exact rate each provider publishes. Image, video, and speech bill per unit of output. The platform subscription is the only thing we charge for, and the free tier covers every model flagged free, forever.
Prices
Per token, per provider rate, zero markup
Input and output rates per one million tokens, exactly as each provider publishes them. What the catalog lists is what the invoice shows.
| Model | Provider | Input / 1M | Output / 1M | Context | Free tier |
|---|---|---|---|---|---|
| Helios 5 | Solstice Labs | $1.25 | $10.00 | 400k | - |
| Helios 5 Mini | Solstice Labs | $0.25 | $2.00 | 400k | - |
| Meridian Pro 4.5 | Meridian AI | $3.00 | $15.00 | 200k | - |
| Meridian Max 4.1 | Meridian AI | $15.00 | $75.00 | 200k | - |
| Tycho 2.5 Deep | Tycho Systems | $1.25 | $10.00 | 1M | - |
| Tycho 2.5 Swift | Tycho Systems | $0.30 | $2.50 | 1M | - |
| Foghorn 4 Coastal free | Foghorn AI | $0.22 | $0.85 | 1M | included |
| Northbeam V3.1 free | Northbeam | $0.56 | $1.68 | 128k | included |
| Vela 3 235B free | Vela Labs | $0.20 | $0.60 | 256k | included |
| Aurora Large 3 | Aurora Compute | $2.00 | $6.00 | 256k | - |
| Beacon A | Lighthouse Labs | $2.50 | $10.00 | 256k | - |
| Pelagic 70B free | Pelagic AI | $0.12 | $0.30 | 128k | included |
Unit prices
Non-token modalities, per unit of output
Images bill per generation, video per second of output, transcription per audio minute, voices per million characters.
| Modality | Unit | From | Models |
|---|---|---|---|
| Image generation | per image | from $0.02 | Tidal Sketch 1, Tidal 1.1 Pro |
| Video generation | per second of output | from $0.35 | Maelstrom 3 |
| Speech to text | per audio minute | from $0.006 | Echo large-v3 |
| Text to speech | per 1M characters | from $15 | Siren Multilingual v2 |
Free tier
Free means free, not a countdown
Every model flagged free in the catalog is free to call on any key, at any volume the rate limits allow. No credit card, no trial clock, no expiry, and the same smart routing and fallbacks as paid traffic.
Upgrade is $5 of credit on the same key.
- Cost
- $0, no credit card, no trial clock
- Coverage
- Every model flagged free in the catalog
- Rate
- 50 requests per minute, 4 concurrent
- Routing
- Same smart routing and fallbacks as paid keys
- Analytics
- Full usage and cost dashboard included
- Upgrade
- Add $5 of credit, keep the same key
Analytics
Watch the spend before it happens
The usage endpoint returns tokens, spend, and latency per key, model, and day. Set hard budgets per key: when a cap is hit, requests stop instead of the invoice growing. Alerts fire at 50, 80, and 100 percent.
curl "https://api.liquidwhale.com/v1/usage?group_by=model&period=30d" \
-H "Authorization: Bearer $LIQUIDWHALE_KEY"FAQ
Questions teams ask before switching
-
We buy tokens at the rate each provider publishes and pass that rate through. Our revenue comes from the platform subscription, not from a hidden percentage on your bill.
-
Every model flagged free in the catalog is free to call, with rate limits at 50 requests per minute and 4 concurrent requests. No credit card, no trial clock, no expiry.
-
No. One Liquid Whale key reaches every provider in the catalog. You never see a provider console unless you want to.
-
The routing layer opens a circuit around it and retries your request on the next healthy provider serving the same model class. The response headers show every attempt.
-
Yes. Set a hard budget per key. When the cap is reached, requests stop instead of surprising you, and alerts go out at 50, 80, and 100 percent.
-
Most teams change the base URL, swap the key, and rerun their eval suite. The request shape stays the same, so the diff is usually two lines.
Get started
Start on the free tier, scale when it works
Create a key, call the free models, and watch the numbers in the usage dashboard. Add credit on the same key when you are ready. Nothing to migrate later.