Billing#
Ultralytics Platform uses credits for metered cloud training. Add credits, track
usage, and manage your subscription from Settings > Billing.

Plans#
Choose the plan that fits your needs. Compare plans in Settings > Plans:

| Feature | Free | Pro ($29/mo) | Enterprise |
|---|---|---|---|
| Signup Credit | $5 / $25* | - | Custom |
| Monthly Credit | - | $30/seat/month | Custom |
| Models | 100 | 500 | Unlimited |
| Concurrent Trainings | 3 | 10 | Unlimited |
| Storage | 100 GB | 500 GB | Unlimited |
Dataset Upload (ZIP/TAR incl. .tar.gz/.tgz/NDJSON) | 10 GB | 20 GB | 50 GB |
| Deployments | 3 | 10 | Unlimited |
| Cloud GPU Types | 24 | 26 | 26 |
| Best GPUs (B200, B300) | - | Yes | Yes |
| Teams | - | Up to 5 members | Custom size |
| SSO / SAML | - | - | Yes |
| Enterprise License | - | - | Yes |
| License | AGPL-3.0 | AGPL-3.0 | Enterprise |
*Free plan: $5 at signup, or $25 if you verify a company/work email address.
Free Plan#
Get started at no cost:
- $5 signup credit ($25 for verified company/work emails)
- Unlimited public and private projects and datasets
- 100 models
- 3 concurrent cloud trainings
- 3 deployments
- 100 GB storage · 10 GB dataset upload limit
- Model export to all 20 formats
- Draw and Smart annotation modes
- 24 cloud GPU types including 5090, H100 & H200 ($0.24–$4.39/hr)
- Community support
Sign up with a company email address (not gmail.com, outlook.com, etc.) to receive $25 in signup credits instead of $5.
Pro Plan#
For professionals and small teams ($29/month or $290/year):
- $30/seat/month in credits (recurring)
- 500 models
- 10 concurrent cloud trainings
- 500 GB storage · 20 GB dataset upload limit
- 10 cloud deployments
- Team collaboration with 4-role RBAC (up to 5 members)
- Access to the best GPUs (B200, B300)
- Priority support
Choose yearly billing ($290/year) to save 17% compared to monthly billing.
Enterprise#
For organizations with advanced needs:
- Custom credit allocation
- Unlimited models, storage, trainings, and deployments · 50 GB dataset upload limit
- Enterprise License (commercial use, non-AGPL)
- SSO / SAML authentication
- On Premise data and compute
- ISO/IEC 27001:2022 and SOC 2 Type I compliance
- Enterprise SLA guarantees
- Enterprise support
See Ultralytics Licensing for Enterprise plan details.
Credits#
Credits are the currency for Platform compute services.
Credit Balance#
View your balance in Settings > Billing:

| Balance Type | Description |
|---|---|
| Available | Credits available for cloud training |
Credit Uses#
Credits are consumed by:
| Service | Rate |
|---|---|
| Cloud Training | GPU rate x hours |
Add Credits#
Top up your balance:
- Go to Settings > Billing
- Click Top Up
- Select or enter amount ($5 - $1,000)
- Complete payment

Top-Up Presets#
| Amount |
|---|
| $10 |
| $20 |
| $50 |
| $100 |
| $500 |
Custom amounts between $5 and $1,000 are also supported.
Auto Top-Up#
Enable automatic credit purchases when your balance drops below a threshold:
- Go to Settings > Billing
- Toggle Auto Top-Up on
- Set Threshold (balance level that triggers a top-up)
- Set Amount (credits to purchase when triggered)
- Click Save
Default settings: threshold $20, amount $100.
Auto top-up can reduce the chance that a paid cloud training job is stopped for insufficient credits. It requires a valid default payment method.
Payment Methods#
Manage payment methods in Settings > Billing:
- Add Card: Click Add Card to add a credit or debit card
- Set as Default: Set a default payment method for top-ups and subscriptions
- Remove: Remove payment methods you no longer need
Billing Address#
Set a billing address for invoices:
- Go to Settings > Billing
- Click Add Address (or Edit if already set)
- Enter your billing details (name, address, country)
- Click Save
Training Cost Flow#
Cloud training estimates cost before start and charges for actual GPU time used.
flowchart LR
A[Start Training]:::start --> B[Estimate Cost]:::proc
B --> C[Run Training]:::proc
C --> D[Charge Actual Usage]:::out
classDef start fill:#4CAF50,color:#fff
classDef proc fill:#2196F3,color:#fff
classDef out fill:#9C27B0,color:#fffHow It Works#
- Estimate: Platform calculates estimated cost based on model size, dataset size, epochs, and GPU
- Authorize Start: Your available balance is checked before training starts
- Train: Job runs on the selected GPU
- Charge: On completion (or cancellation), billing uses actual runtime
You pay for actual compute time used, including partial runs that are cancelled.
Training Costs#
Cloud training costs depend on GPU selection:
| GPU | Generation | VRAM | Cost/Hour | Best For |
|---|---|---|---|---|
| RTX 2000 Ada | Ada | 16 GB | $0.24 | Small datasets, testing |
| RTX A4500 | Ampere | 20 GB | $0.25 | Small-medium datasets |
| RTX 4000 Ada | Ada | 20 GB | $0.26 | Medium datasets |
| RTX A5000 | Ampere | 24 GB | $0.27 | Medium datasets |
| L4 | Ada | 24 GB | $0.39 | Inference optimized |
| A40 | Ampere | 48 GB | $0.44 | Larger batch sizes |
| RTX 3090 | Ampere | 24 GB | $0.46 | General training |
| RTX A6000 | Ampere | 48 GB | $0.49 | Large models |
| RTX PRO 4000 | Blackwell | 24 GB | $0.57 | Budget Blackwell |
| RTX PRO 4500 | Blackwell | 32 GB | $0.64 | Great price/performance |
| RTX 4090 | Ada | 24 GB | $0.69 | Best price/performance |
| RTX 6000 Ada | Ada | 48 GB | $0.77 | Large batch training |
| L40S | Ada | 48 GB | $0.86 | Large batch training |
| RTX PRO 5000 | Blackwell | 48 GB | $0.96 | Large batch training |
| RTX 5090 | Blackwell | 32 GB | $0.99 | Latest consumer generation |
| L40 | Ada | 48 GB | $0.99 | Large models |
| A100 PCIe | Ampere | 80 GB | $1.39 | Production training |
| A100 SXM | Ampere | 80 GB | $1.49 | Production training |
| RTX PRO 6000 | Blackwell | 96 GB | $2.09 | Recommended default |
| H100 PCIe | Hopper | 80 GB | $2.89 | High-performance training |
| H100 NVL | Hopper | 94 GB | $3.19 | Maximum performance |
| H100 SXM | Hopper | 80 GB | $3.29 | Fastest training |
| H200 NVL | Hopper | 143 GB | $3.39 | Maximum memory |
| H200 SXM | Hopper | 141 GB | $4.39 | Maximum performance |
| B200 | Blackwell | 180 GB | $5.89 | Large models (Pro+) |
| B300 | Blackwell | 288 GB | $7.39 | Largest models (Pro+) |
B200 and B300 GPUs require a Pro or Enterprise plan. All other GPUs are available on all plans.
See Cloud Training for complete GPU options and pricing.
Cost Calculation#
Total Cost = GPU Rate x Training Time (hours)Example: Training for 2.5 hours on RTX PRO 6000
$2.09 x 2.5 = $5.23Upgrade to Pro#
Upgrade for more features and monthly credits:
- Go to Settings > Plans
- Click Upgrade to Pro
- Choose billing cycle (Monthly or Yearly)
- Complete checkout

Pro Benefits#
After upgrading:
- $30/seat/month credit added immediately and each month
- Storage increased to 500 GB · 20 GB dataset upload limit
- 500 models
- 10 concurrent cloud trainings
- 10 cloud deployments
- Team collaboration (up to 5 members)
- Access to best GPUs (B200, B300)
- Priority support
Cancel Pro#
Cancel anytime from the Plans tab:
- Go to Settings > Plans
- Click Cancel Subscription on the Pro plan card
- Confirm in the dialog
If you cancel before the end of your billing period, a Resume Subscription button appears — click it to undo the cancellation before the period ends.
Pro features remain active until the end of your current billing period. Monthly credits stop being granted at cancellation.
Downgrading to Free#
When your Pro subscription ends (cancelled or expired), your account reverts to the Free plan. Here's what happens to your existing resources:
| Resource | What Happens |
|---|---|
| Models | All models preserved. Cannot create new models beyond 100-model limit |
| Deployments | All deployments preserved. Cannot create new beyond 3-deployment limit |
| Storage | All data preserved. Cannot upload new data beyond 100 GB limit |
Dataset Upload (ZIP/TAR incl. .tar.gz/.tgz/NDJSON) | Upload limit reduced from 20 GB to 10 GB per file |
| Credit Balance | Existing credits preserved and usable |
| Monthly Credits | $30/seat/month grants stop immediately |
| Team Members | Members notified and lose access to team resources |
| GPU Access | Standard GPUs remain available. Best GPUs (B200, B300) require Pro or Enterprise |
| Concurrent Trainings | Limit reduced from 10 to 3 |
Downgrading does not automatically delete models, datasets, or deployments. The workspace owner retains access, while Free-plan creation limits apply and team members lose access to team resources.
Transaction History#
View all transactions in Settings > Billing:

| Column | Description |
|---|---|
| Date | Transaction date |
| Type | Signup, Purchase, Subscription, Monthly Grant, Training, Refund, Adjustment, Promo, Auto Top-Up, Auto Top-Up Failed, Pro Credit Expiry |
| Amount | Transaction value (green for credits, red for charges) |
| Balance | Resulting balance after transaction |
| Details | Additional context (model link, receipt, period) |
FAQ#
- Running paid cloud training: Stops when metered usage pushes the balance below zero
- New training: Cannot start new jobs until balance is positive
- Deployments: Continue running regardless of balance
Add credits to restore a positive balance before starting new training jobs. Enable auto top-up to reduce the chance of an active job being stopped for insufficient funds.
Transaction receipts are available in the transaction history. Click the receipt icon next to any purchase transaction.
If a cloud GPU has started, failed, cancelled, completed, and auto-terminated jobs are charged for elapsed GPU time. Validation or launch failures before cloud compute starts have no GPU usage charge. See Cloud Training Billing for the full breakdown.
The Free plan includes $5 in signup credit, increased to $25 after verifying a company email. No credit card is required to start.