- Pricing
Credit pricing
Model-specific packages
Choose an exact model SKU. Each package has a separate quota, not a shared family balance. Model credits are used first for matching requests, then your all-in-one balance is the fallback. Daily Unlimited coverage remains unchanged.
deepseek-v4-pro
200M credits · Rp25.000
This exact SKU only · excludes provider fees
deepseek-v4.1-flash
400M credits · Rp25.000
This exact SKU only · excludes provider fees
glm-5.2
200M credits · Rp25.000
This exact SKU only · excludes provider fees
glm-5.3
200M credits · Rp25.000
This exact SKU only · excludes provider fees
glm-5.3-flash
400M credits · Rp25.000
This exact SKU only · excludes provider fees
All-in-one prepaid credits
No subscription. All catalog models draw from one balance, settled after each request. Unused credits stay on your key.
More credit amounts · save up to 5%
Starter packages have different credit-per-rupiah rates. Full packs contain 100M credits, with bulk discounts up to 5%.
Credits are added after payment is verified.
Daily Unlimited
Time-limited access, not a credit top-up. GLM · DeepSeek · MiniMax only; other models continue to use your regular credits.
Daily Unlimited is temporarily not on sale. Prepaid credits keep working exactly as before, and GLM · DeepSeek · MiniMax usage simply draws from your balance.
QRIS checkout in your workspace. Provider fees are shown separately before payment.
Model rates
Rates are in credits, not rupiah. The multiplier is the credit rate per input token. Cached input and output use model-specific weights. Rates below come from the gateway manifest; models without a rate are pending pricing.
18 models · Billing: credit · Settled per request
Claude 1 models
| Model | Input / token | Cached / token | Output / token | Minimum / request |
|---|---|---|---|---|
| claude-opus-4.7 | 10 | 5 | 50 | 1,000 |
Deepseek 4 models
| Model | Input / token | Cached / token | Output / token | Minimum / request |
|---|---|---|---|---|
| deepseek-v4-mod | 6 | 1.8 | 30 | 1,000 |
| deepseek-v4-pro | 3 | 1.5 | 15 | 1,000 |
| deepseek-v4.1-flash | 0.75 | 0.375 | 3.75 | 1,000 |
| deepseek-v4.1-flash-1017-mod | 6 | 3 | 30 | 1,000 |
Glm 7 models
| Model | Input / token | Cached / token | Output / token | Minimum / request |
|---|---|---|---|---|
| glm-5.2 | 2.5 | 1.25 | 10 | 1,000 |
| glm-5.2-mod | 6 | 3 | 24 | 1,000 |
| glm-5.3 | 2.5 | 0.75 | 10 | 1,000 |
| glm-5.3-flash | 2 | 1 | 10 | 1,000 |
| glm-5.3-flash-mod | 6 | 3 | 30 | 1,000 |
| glm-5.3-flashx-mod | 6 | 3 | 30 | 1,000 |
| glm-5.3-mod | 6 | 3 | 30 | 1,000 |
Image 4 models
| Model | Input / token | Cached / token | Output / token | Minimum / request |
|---|---|---|---|---|
| image-gemini-2.5-flash-image | 2 | 1 | 10 | 1,100 |
| image-gemini-3.0-pro-image | 2 | 1 | 10 | 1,100 |
| image-gemini-3.1-flash-image | 2 | 1 | 10 | 1,100 |
| image-gpt-image-2 | 2 | 1 | 10 | 1,100 |
Kimi 2 models
| Model | Input / token | Cached / token | Output / token | Minimum / request |
|---|---|---|---|---|
| kimi-k3 | 4 | 2 | 16 | 1,000 |
| kimi-k3-mod | 6 | 3 | 24 | 1,000 |
How credits are calculated
Each request is settled after it completes. Input, cached input, and output use the model's rates and weights, with a minimum per-request charge.
Formula
Applied to every request, then rounded up.
credits = ceil(input*mult + cache*mult*cache_weight + output*mult*output_weight*surge_multiplier), min = floor_credits- mult
- — per-model multiplier (rate per input token)
- cw
- — cache weight, the discounted fraction for cached input
- ow
- — output weight, the multiple applied to output tokens
- surge
- — extra output multiplier during peak demand, when set
- floor
- — minimum charge per request (typically 1,000 credits)
Worked example
deepseek-v4-pro — mult ×3, floor 1,000
Example without an additional surge multiplier.
- 1,000 input × 3
- 3,000
- 500 output × 3 × 5
- 7,500
- 2,000 cached × 3 × 0.5
- 3,000
- Charged
- 13,500 credits
Ready to connect an application? Read the integration guide.