One API for leading open-weight models
LarkTide API gives developers and teams OpenAI-compatible access to GLM, DeepSeek, Kimi and MiniMax models: one API key, prepaid US-dollar credits, usage tracking per key and spending limits.
Why LarkTide API
One key, many models
Switch between GLM, DeepSeek, Kimi and MiniMax models by changing the model parameter. No separate accounts with each provider.
OpenAI-compatible
Point your existing OpenAI SDK or HTTP client at the LarkTide endpoint and keep your code.
Usage you can see
Every request is metered per API key: model, tokens and cost, in your console.
Spending limits
Set a quota on each key so a runaway script cannot drain your balance.
Pay as you go
Prepaid US-dollar credits at the providers' official per-token rates. No subscription, and credits do not expire.
Your data
For normal requests we do not store your prompts or outputs; error responses from a model provider may be logged for troubleshooting. Requests go to the provider of the model you choose.
Models
The Flash models are fast and low-cost for high-volume work; the flagship models are for harder tasks. Full rates are on the pricing page.
| Model | API id | Tier | Context | Input / Output per 1M tokens |
|---|---|---|---|---|
| GLM-5.3-Flash | glm-5.3-flash | Flash | 1M | $0.15 / $0.50 |
| DeepSeek-V4.1-Flash | deepseek-flash | Flash | 1M | $0.30 / $1.20 |
| Kimi K2.6 | kimi-k2.6 | Flash | 262K | $0.95 / $4.00 |
| MiniMax-M3 | MiniMax-M3 | Flash and flagship | 1M | $0.30 / $1.20 (up to 512K input) |
| GLM-5.3 | glm-5.3 | Flagship | 1M | $1.40 / $4.40 |
| DeepSeek-V4-Pro | deepseek-v4-pro | Flagship | 1M | $1.32 / $3.96 |
| Kimi K3 | kimi-k3 | Flagship | 1M | $3.00 / $15.00 |
Models are developed and served by their providers. LarkTide provides the unified API, API keys, metering and billing.
How it works
- Request accessEmail access@larktide.com with a short note on your use case. We send an invitation code.
- Create your accountSign up in the LarkTide console with your invitation code and create an API key.
- Add creditsBuy prepaid US-dollar credits by card. Payments will be processed by Stripe.
- Call the APIUse the OpenAI-compatible endpoint with your key and the model id you want.
Quick start
curl https://api.larktide.com/v1/chat/completions \
-H "Authorization: Bearer $LARKTIDE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-flash",
"messages": [{"role": "user", "content": "Hello!"}]
}'The endpoint works with API keys created in the console after your account is activated.
Questions
Who can use LarkTide API?
Developers and businesses aged 18 or over, for professional use. During the beta, access is by invitation. The service is not offered in mainland China or in countries and regions subject to comprehensive U.S. sanctions.
How does billing work?
You buy prepaid credits in packs from $10 to $500. A 5% platform fee applies to each purchase, so a $100 purchase adds $95 of credits. Each request is charged at the model's per-token rate. See Pricing and Credits & Refunds.
Do you store my prompts?
Not for normal requests. We keep usage metadata such as model, token counts and cost, not the content of your requests or outputs. If a model provider returns an error, its error response may be logged for troubleshooting. Content is processed by the provider of the model you select. Details are in the Privacy Policy.
Can I get a refund?
Unused paid credits from a purchase can be refunded to the original payment method within 7 days of that purchase. The platform fee and credits already used are not refundable. See Credits & Refunds.