Token Credits & Pricing

Documentation on token credits management, billing structure, and pricing models for Neosantara APIs.

Quick Access

Check your Current Balance or Top Up Credits

Pay-As-You-Go System

Neosantara AI operates on a flexible Pay-As-You-Go billing system. You only pay for what you use—no mandatory subscriptions, no restrictive volume limits.

No Subscriptions

Pay only for actual usage without monthly commitments

Flexible Scaling

Scale up or down based on your needs

Transparent Pricing

Clear per-token pricing with no hidden fees

---

Legacy Users (Transition)

This section applies only to users registered before the Pay-As-You-Go update.

Legacy plan users operate under a special transition model:

1. Quota Usage

   Your monthly token quota remains active and is prioritized for billing

2. Automatic Migration

   When you exceed your monthly quota for the first time, your account automatically migrates to Pay-As-You-Go mode permanently

3. Welcome Credit

   Receive a one-time Rp 10,000 credit upon migration to ensure smooth service continuity

---

Pricing Structure

All prices are in IDR (Indonesian Rupiah). Models with USD pricing are converted using real-time exchange rates.

Batch Processing Discount

Save 50% with Batch API

Process large volumes asynchronously at half the cost of standard API calls. Perfect for non-urgent, high-volume tasks.

Usage Type	Rate Adjustment
Batch Input Tokens	50% of standard input price
Batch Output Tokens	50% of standard output price

The 50% discount applies automatically when using the /v1/batches endpoint. No special configuration needed!

---

Chat Completion Models

Pricing per 1 million tokens (1M tokens):

Indonesian Models

Model	Input/1M	Output/1M	Cache Read/1M	Cache Write/1M
nusantara-base	Rp 300	Rp 1,500	N/A	N/A
garda-beta-mini	Rp 2,500	Rp 10,499	N/A	N/A
archipelago-70b	Rp 4,710	Rp 36,543	N/A	N/A
sea-lion-v4-27b-it	Rp 350	Rp 560	N/A	N/A
gpt-oss-20b	Rp 400	Rp 1,600	N/A	N/A

Open Source Models

Model	Input/1M	Output/1M	Cache Read/1M	Cache Write/1M
llama-3.3-70b-instruct	Rp 4,560	Rp 31,991	N/A	N/A
llama-3.3-70b-turbo	Rp 300	Rp 1,200	N/A	N/A
llama-3.3-nemotron-super-49b-v1.5	Rp 4,560	Rp 31,991	N/A	N/A
llama-3.2-11b	Rp 100	Rp 400	N/A	N/A
gemma2-9b-it	Rp 200	Rp 1,000	N/A	N/A
gemma-3-27b-it	Rp 200	Rp 1,500	N/A	N/A
granite-3-8b-instruct	Rp 2,880	Rp 11,520	N/A	N/A

Specialized Models

Model	Input/1M	Output/1M	Cache Read/1M	Cache Write/1M
deepseek-chat-v3.1	Rp 3,333	Rp 8,333	N/A	N/A
qwen3-coder	Rp 3,333	Rp 8,333	N/A	N/A
kimi-k2	Rp 16,642	Rp 49,929	N/A	N/A
glm-4.6	Rp 823	Rp 2,823	N/A	N/A
glm-4.6-plus	Rp 823	Rp 2,832	N/A	N/A

Premium Models (USD)

Model	Input/1M	Output/1M	Cache Read/1M	Cache Write/1M
claude-4.5-sonnet	$3.00	$15.00	$0.30	$3.75
gemini-3-flash	$0.50	$3.00	$0.10	$0.03
grok-code-fast	$0.20	$1.50	$0.05	$0.13
grok-4.1-fast-non-reasoning	$0.30	$0.72	$0.10	$0.15

USD prices are converted to IDR at checkout using current exchange rates.

Understanding Caching

Cache Read: When a model supports prompt caching, repeated input tokens are read from cache at a lower rate (typically 90% cheaper). If “N/A”, the model doesn’t support caching.

Cache Write: The first time tokens are cached, they’re billed at the standard input rate. Subsequent reads use the cheaper Cache Read rate.

Savings Example: With Claude 4.5 Sonnet, cache reads cost $0.30/1M vs$3.00/1M for standard input—a 90% savings on repeated content!

How Token Pricing Works

Tokens are the basic units of text processing. Roughly:

• 1 token ≈ 4 characters in English
• 1 token ≈ 2-3 characters in Indonesian
• 1M tokens ≈ 750,000 words in English

Example Calculation:

• Model: nusantara-base
• Input: 1,000 tokens (Rp 300 / 1M × 1,000 = Rp 0.3)
• Output: 500 tokens (Rp 1,500 / 1M × 500 = Rp 0.75)
• Total: Rp 1.05 per request

---

Image Generation

High Resolution

1024×1024

Rp 1,500 per image

Medium Resolution

512×512

Rp 800 per image

Low Resolution

256×256

Rp 400 per image

---

OCR (Optical Character Recognition)

Extract text from images at an affordable rate:

Model	Price per Image
deepseek-ocr	Rp 100

OCR pricing is per image, regardless of text length extracted. Process multiple pages efficiently!

---

Tool Usage & Function Calling

Tool usage costs are included in standard token pricing—no additional orchestration fees!

When using function calling or tools:

• Tool definitions count as input tokens
• Tool calls count as output tokens
• Tool results count as input tokens

All billed at your selected model’s standard rate.

Example: Using tools with nusantara-base:

• Tool definition: 200 tokens input = Rp 0.06
• Model’s tool call: 50 tokens output = Rp 0.075
• Tool result: 150 tokens input = Rp 0.045
• Total tool overhead: Rp 0.18

---

Throughput Tiers

Automatic Tier Upgrades

Your tier automatically upgrades as your cumulative lifetime deposits increase. No manual action needed!

Rate Limits by Tier

Tier	Min. Cumulative Deposit	RPM	ITPM	OTPM	Batch API
Free	Rp 0	3	5,000	2,000	❌
Basic	Rp 85,000	50	20,000	5,000	✅
Standard	Rp 670,000	1,000	100,000	25,000	✅
Pro	Rp 3,350,000	2,000	200,000	50,000	✅
Enterprise	Rp 6,700,000	4,000	500,000	125,000	✅

What do RPM, ITPM, and OTPM mean?

• RPM (Requests Per Minute): Maximum API requests you can make per minute
• ITPM (Input Tokens Per Minute): Maximum input tokens processed per minute across all requests
• OTPM (Output Tokens Per Minute): Maximum output tokens generated per minute across all requests

All three limits apply simultaneously. Exceeding any limit will result in rate limit errors (429).

How tier upgrades work

Tiers are based on your total lifetime deposits, not current balance:

Example:

1. You deposit Rp 100,000 → Unlock Basic tier
2. You spend Rp 50,000 → Still Basic tier (deposit history counts)
3. You deposit another Rp 600,000 → Unlock Standard tier

Once unlocked, you never lose tier benefits, even if your balance decreases.

Need higher limits?

For custom rate limits beyond Enterprise tier, contact our sales team to discuss:

• Custom concurrent request limits
• Dedicated infrastructure
• Volume discounts
• Priority support

---

Balance Management

Checking Your Balance

Usage Dashboard

View real-time balance and usage statistics

Billing Page

Top up credits and manage payment methods

Adding Funds

1. Navigate to Billing

   Go to the Billing Page in your dashboard

2. Select Top Up Amount

   Click “Top Up Balance” and choose your desired amount

3. Complete Payment

   Pay via QRIS, GoPay, OVO, or Bank Transfer. Credits are added instantly!

Supported Payment Methods

QRIS

Instant payment via any QRIS-enabled app

E-Wallets

GoPay, OVO, and other supported e-wallets

Bank Transfer

Direct transfer from Indonesian banks

Credit Card

Contact support for credit card payments

For credit card payments or large enterprise deposits, please contact us at [e-mail]

---

Cost Optimization Tips

Use Batch API for 50% Savings

For non-urgent tasks, always use the Batch API to cut costs in half. Perfect for:

• Data labeling and classification
• Bulk translations
• Content generation pipelines
• Embeddings for large datasets

Choose the Right Model

Match model capability to task complexity:

• Simple tasks: Use llama-3.2-11b (Rp 100/1M) or nusantara-base (Rp 300/1M)
• Medium tasks: Use gpt-oss-20b (Rp 1,600/1M) or gemma2-9b-it (Rp 200/1M)
• Complex tasks: Use claude-sonnet-4.5 or specialized models only when needed

Savings: Using a smaller model can reduce costs by 80-95%!

Leverage Prompt Caching

For models supporting caching (like Claude), structure prompts to reuse content:

• Put static instructions/examples at the start
• Cache reads cost 90% less than fresh inputs
• Ideal for repeated queries with consistent context

Optimize Token Usage

• Set appropriate max_tokens limits
• Use concise system prompts
• Implement streaming to stop generation early if needed
• Remove unnecessary whitespace and formatting

---

Frequently Asked Questions

What happens if my balance runs out?

API requests will fail with a 402 (Payment Required) error. Top up your balance to resume service. No data is lost during service interruption.

Are there any hidden fees?

No. You only pay for tokens consumed and images generated. No setup fees, maintenance fees, or API access charges.

Can I get a refund?

Unused balance can be refunded within 30 days of deposit, minus payment processing fees (2-3%). Contact support for refund requests.

Do credits expire?

No. Your balance never expires as long as your account remains active (at least one API call per 12 months).

How accurate is the usage tracking?

Usage is tracked in real-time with millisecond precision. You can view detailed breakdowns per request in your dashboard.

Can I set spending limits?

This feature is not yet implemented. However, since Neosantara uses a prepaid credit system, you’re never automatically charged—you can only spend up to your available balance.

Need Help?

Have questions about pricing or billing? Contact our support team at [e-mail]