Pricing & Credits

Understand MLStack's pricing model and credit system.

Credit System

MLStack uses a pay-as-you-go credit system:

Purchase credits upfront
Charged per inference request
No monthly fees or subscriptions
Only pay for what you use

Pricing

Inference Costs

Inference costs vary based on model size:

Small models (1B-3B): ~$0.0044 per request
Medium models (7B-14B): ~$0.0264 per request

Key benefits:

Charged per request - Only pay for actual inference
No idle costs - No charges when not using the service
Transparent - Cost calculated from GPU time

Agent Execution

Agent costs = LLM calls + tool calls:

Each iteration may call LLM (charged based on model size)
Tool calls are free (external APIs may have own costs)
Max iterations limits cost

Example (8B model):

Agent with max 5 iterations:
- Worst case: 5 LLM calls = ~$0.022
- Typical: 2-3 calls = ~$0.009-$0.013

Example (14B model):

Agent with max 5 iterations:
- Worst case: 5 LLM calls = ~$0.132
- Typical: 2-3 calls = ~$0.053-$0.079

Free Credits

New users receive $5 in free credits:

Granted on signup
Expires after 30 days
~1,100 requests (1B-3B models) or ~190 requests (7B-14B models)

Checking Balance

View your balance on the Billing page:

Current Balance: $42.50

Adding Credits

Go to Billing
Select a credit pack under Purchase Credits ($10, $25, $50, or $100)
Complete checkout via Stripe
Credits are instantly available

Cost Optimization

Batch Requests

Submit multiple jobs efficiently:

# Submit all at once
job_ids = [submit(p) for p in prompts]
# Poll together
results = [poll(jid) for jid in job_ids]

Cache Results

Cache common queries to avoid re-inference:

import hashlib
cache = {}

def get_or_infer(prompt):
    key = hashlib.md5(prompt.encode()).hexdigest()

    if key in cache:
        return cache[key]

    result = submit_and_poll(prompt)
    cache[key] = result
    return result

Use Structured Output

Define schemas to get exactly what you need without extra tokens.

Monitor Usage

Track spending in real-time:

Daily/weekly email summaries
Spending alerts
Usage analytics

Billing

No surprise charges
Prepaid model - can't overspend
Detailed transaction history
Export invoices for accounting

Enterprise Pricing

Need volume discounts? Contact [email protected] for:

Custom pricing tiers
Committed use discounts
Dedicated support
SLA guarantees

Next Steps

Getting Started - Quick start tutorial
API Reference - Complete endpoint documentation

Documentation