Pricing & Billing
TK-NovaLink offers transparent and flexible billing with pay-as-you-go and per-call pricing models.
Billing Models
We support two billing models:
| Billing Mode | Description |
|---|---|
| Pay-as-you-go | Billed by token count (input and output tokens priced separately) |
| Per-use billing | Billed per call / image / second; applies to image and video generation, etc. |
When a model supports both modes, per-use billing takes priority by default. If a token is set to "pay-as-you-go only", token-based billing is enforced.
Token Billing Basics
Cost = Input Tokens × Input Price + Output Tokens × Output Price. 1 Chinese character ≈ 2 tokens.
| Scenario | Input Tokens | Output Tokens | Total |
|---|---|---|---|
| Simple Q&A | 20 | 50 | 70 |
| Article writing | 100 | 2000 | 2100 |
| Code debugging | 500 | 300 | 800 |
| Long-form translation | 5000 | 5000 | 10000 |
Popular Models
Here are our recommended popular models:
| Model ID | Positioning | Best For |
|---|---|---|
| gpt-4o | Flagship GPT model | General chat, complex reasoning |
| gpt-4o-mini | Lightweight model | Classification, extraction, batch processing |
| claude-sonnet-4 | Coding workhorse | Coding agents, long-running tasks |
| gemini-2.5-pro | Balanced performance and cost | General chat, multimodal |
| deepseek-v3 | High cost-performance | General chat, code generation |
Cost Estimation
We recommend testing with a small sample, checking token consumption in the console, and calculating total cost.
For exact model IDs, token groups, billing modes, and actual unit prices, refer to the real-time pricing in the console. Do not estimate bills from upstream providers’ list prices; rely on the console and call logs.
Enterprise Services
For large purchases and enterprise customers, we provide dedicated services:
| Service | Description |
|---|---|
| Dedicated account manager | One-on-one technical support |
| SLA guarantee | 99.99% availability |
| Custom invoicing | IT service fee invoice |
| Contact | support@tk-novalink.com |
Next
FAQ