Credits & Earnings
How hosting earns you credits, how they're calculated, and how to spend them on other models.
Credits & Earnings Guide
Everything you need to know about how hosting earns you credits and how to spend them.
💰 How You Earn
Every request your server handles earns you credits. You set your own price per token when you register your models.
Platform fee = 20% of your earning (kept by NeuralGate)
Customer pays = your listed price (unchanged) Hoster earns = your listed price × 80% Platform fee = your listed price × 20%
Example: Serving a 500-token request
| Item | Calculation | Amount |
|---|---|---|
| Input tokens | 250 tokens × $0.10/1M | $0.000025 |
| Output tokens | 250 tokens × $0.30/1M | $0.000075 |
| Customer pays | your listed price | $0.000100 |
| Platform fee (20%) | $0.000100 × 20% | $0.000020 |
| Your net earning | customer price × 80% | $0.000080 |
🏷️ Setting Your Price
Prices are in microdollars per 1 million tokens. 1 USD = 1,000,000 microdollars.
| Price field value | Effective rate | Good for |
|---|---|---|
50 | $0.05 per 1M tokens | Small models (0.5B–2B), high volume |
100 | $0.10 per 1M tokens | 7B–9B models, competitive rate |
300 | $0.30 per 1M tokens | Standard output pricing for small models |
500 | $0.50 per 1M tokens | 13B–27B models |
2000 | $2.00 per 1M tokens | 70B+ models, premium quality |
OpenRouter's current rates for comparison: Qwen 3.5 9B at $0.05/1M input · Gemma 4 26B at $0.05/1M input (free tier available). Pricing competitively increases your traffic share.
⏱️ Credit Lifecycle
Request served immediate
Your server handles a request. Tokens are counted and your earning is calculated.
Pending period 24 hours
Earnings are held as pending for 24 hours. This prevents fraud and allows time for dispute resolution. Your pending balance is visible in the portal but not yet spendable.
Credits released spendable
After 24h, pending credits move to your linked customer account and become spendable. You can immediately use them to query any other model on NeuralGate.
Spend or cash out
Use your API key to call other models — credits are deducted automatically. Cash payouts are available once your balance exceeds $10 (coming soon via Stripe).
📊 Checking Your Balance
Three ways to see your earnings:
1. Hoster Portal (easiest)
Go to /hosters/portal → see your total earnings, pending credits, and spendable balance at a glance.
2. API — Earnings statement
Use your linked API key (the one generated when you registered):
curl https://api.computeshare.servequake.com/v1/earnings?days=30 \ -H "Authorization: Bearer ngk_your_hoster_key"
Returns daily breakdown, total earned, and per-model earnings.
3. API — Earnings by model
curl https://api.computeshare.servequake.com/v1/earnings/by-model?days=30 \ -H "Authorization: Bearer ngk_your_hoster_key"
See which of your models earns the most.
🔄 Spending Your Credits
Your hoster account is linked to a customer account. Every time you earn credits, they flow into your customer balance (after the 24h hold). Use the same API key to access other models:
from openai import OpenAI
# Use your hoster API key — credits come from your earnings
client = OpenAI(
api_key="ngk_your_hoster_key",
base_url="https://api.computeshare.servequake.com/v1"
)
# Query any model — debited from your earned credits
response = client.chat.completions.create(
model="auto",
messages=[{"role": "user", "content": "Summarize this document..."}]
)
print(response.choices[0].message.content)
Check your spendable balance
curl https://api.computeshare.servequake.com/v1/usage \ -H "Authorization: Bearer ngk_your_hoster_key" # Returns: credit_balance_usd (your spendable balance)
📈 What Affects How Much Traffic (and Earnings) You Get
NeuralGate routes requests to the best available hoster. Your share of traffic depends on:
| Factor | Impact | How to improve |
|---|---|---|
| Trust score | High — routes prefer trust > 0.8 | Keep server reliable and responsive |
| Health score | High — drops if health checks fail | Ensure /health returns 200 |
| Consistency score | High — tested every 10 min | Serve correct answers to factual questions |
| Latency | Medium — faster = more traffic | Optimize GPU layers, use quantized models |
| Price | Medium — cheapest tier prefers you | Price competitively vs other hosters |
| Moderation status | Critical — probation = 50% traffic, suspended = 0% | Don't violate ToS, respond to warnings |
💳 Cash Payouts (Coming Soon)
Credit-to-cash payouts via Stripe are coming soon. When available:
- Minimum payout threshold: $10.00
- Payment methods: PayPal, bank transfer, or crypto (TBD)
- Payout frequency: on request, processed within 3–5 business days
- Platform fee: 20% already deducted from earnings (no additional fee at payout)
Until then, all earnings are in NeuralGate credits redeemable on any model in the network.
❓ FAQ
Why is my earning showing as "pending"?
All earnings are held for 24 hours before becoming spendable. This protects against fraud and allows dispute resolution. Check your portal — it shows both pending and available balances separately.
I'm serving traffic but earning $0 — why?
Most likely your requests are coming through the free/local routing tier (model="auto" with simple queries). The router may be serving them locally rather than through your hoster endpoint. Check your hoster portal for actual request counts routed to you.
Can I set different prices for different models?
Yes — each model you register has its own price_per_input_token and price_per_output_token. Larger models typically earn more per token.
What's the 20% platform fee?
NeuralGate takes 20% of your gross earnings to cover infrastructure, routing, trust monitoring, and platform operations. You keep 80%. The customer pays your price + 20%.
How do I rotate my bearer token?
Go to your hoster portal and use the "Rotate Token" button. Your server must be updated with the new token immediately — the old one is invalidated instantly.
Start earning today
Register your GPU server and start receiving traffic in minutes.
Register as a Hoster → My Portal →