Two billing modes designed for different usage patterns. No hidden fees, no surprises.
Prepaid credits for developers who want simplicity
Bring your own key for maximum control and scale
| Tier | Amount | Credits Received | Fee | Bonus |
|---|---|---|---|---|
| Starter | $20 | $18.90 | $0.55 (5.5%) | -- |
| Standard | $50 | $47.25 | $2.75 (5.5%) | -- |
| Professional | $100 | $94.50 | $5.50 (5.5%) | BYOK limit increased to 2M |
| Enterprise | Custom | Custom | Negotiated | Dedicated support, SLA |
Per-token prices in USD. Data fetched from API in real-time.
| Model | Provider | Input ($/1M tokens) | Output ($/1M tokens) |
|---|---|---|---|
| Loading... | |||
Standard mode: You pay provider_cost x (1 + 5.5%). The 5.5% covers payment processing and gateway fees.
BYOK mode: You pay nothing for the first 1M requests per month. After that, you pay provider_cost x 5% as a routing fee.
In both cases, you pay only for what you use. No monthly subscription fees.
Yes. You can switch between modes per request by including or omitting the provider_key field. Both modes are available on the same account.
We accept major credit and debit cards via Stripe. Cryptocurrency (USDC) support is coming in Q3 2026.
Credits expire 365 days from your last account activity. Inactive accounts are notified via email before expiration.
Every API request counts as one request regardless of token count. The free tier (1M/month) is shared across all models. The counter resets on the 1st of each month at 00:00 UTC.
Unused credits are refundable within 14 days of purchase. Please contact our support team to initiate a refund.
DeepSeek, GLM, Qwen, MiniMax, Kimi — Chinese model tokens cost 60-99% less than GPT-4 or Claude. One OpenAI-compatible API. Pay with USD via Stripe. No Chinese bank account needed.
No credit card required. Activation link sent to your email.
// Access the cheapest AI tokens on earth // One line change from OpenAI SDK const openai = require('openai'); const client = new OpenAI({ baseURL: "https://api.find-token.com/v1", apiKey: "sk-your-key-here" }); // deepseek-chat: $0.14/M input tokens // vs GPT-4o: $2.50/M input tokens // Save 94% on every request const result = await client.chat.completions.create({ model: "deepseek-chat", messages: [ { role: "system", content: "You are a helpful assistant." }, { role: "user", content: "Write efficient Go code" } ] });
Price per 1 million tokens in USD. Chinese models deliver the same quality at a fraction of the cost.
| Provider | Model | Input ($/1M) | Output ($/1M) | vs GPT-4o |
|---|---|---|---|---|
| OpenAI | GPT-4o | $2.50 | $10.00 | -- |
| Anthropic | Claude 3.5 Sonnet | $3.00 | $15.00 | -- |
| Loading model prices... | ||||
Prices in USD per 1M tokens. Data from API in real-time. A 5.5% gateway fee applies in Standard mode.
Prices in USD per 1M tokens. Updated weekly. A 5.5% gateway fee applies in Standard mode.
Chinese AI models now rival GPT-4 in benchmarks. The only difference is the price tag.
DeepSeek at $0.14/M input vs GPT-4o at $2.50/M. That's 94% savings. At scale, this means thousands of dollars back in your pocket every month.
Drop-in replacement for OpenAI SDK. Change the base URL and your existing code works. Same interface, same tooling, dramatically cheaper tokens.
No Alipay, no WeChat, no Chinese bank account. Use your international credit card. $10 minimum recharge. Transparent 5.5% fee, no hidden costs.
Standard mode: prepay with credits at 5.5% fee. BYOK mode: bring your own provider key, first 1M requests/month free, then 5% routing fee.
Auto-route to cheapest or healthiest provider. If one model goes down, requests automatically fall back to alternatives. 99.9% uptime across all models.
DeepSeek, GLM, Qwen, MiniMax, Kimi, Stepfun — all accessible through a single endpoint. Add new models with zero code changes on your side.
For a typical app processing 10M input + 5M output tokens per month:
| Service | Model | Monthly Cost | Annual Cost | Savings vs GPT-4o |
|---|---|---|---|---|
| OpenAI | GPT-4o | $75,000 | $900,000 | -- |
| DeepSeek (via find-token) | deepseek-chat | $2,800 | $33,600 | 96% less |
| Zhipu (via find-token) | glm-4-flash | $2,000 | $24,000 | 97% less |
| Alibaba (via find-token) | qwen-turbo | $2,000 | $24,000 | 97% less |
| Stepfun (via find-token) | step-3-flash | $1,600 | $19,200 | 98% less |
Every major Chinese model provider. Priced per 1M tokens in USD. No markups, no surprises.
No credit card required. Just an email address.
Sign up with just your email. We'll send you an activation link instantly.
Once activated, create your first API key from the dashboard.
Update your OpenAI client base URL. Your existing code works unchanged.
Start using Chinese model tokens at a fraction of GPT-4 pricing.
Choose what works for your scale. Both give you access to the same cheap Chinese tokens.
Prepaid credits — best for individuals and small teams
Bring your own provider key — zero gateway costs up to 1M/month
Join developers saving 80-97% on AI inference costs. Same quality, fraction of the price.
Get Started FreeTwo billing modes designed for different usage patterns. No hidden fees, no surprises.
Prepaid credits for developers who want simplicity
Bring your own key for maximum control and scale
| Tier | Amount | Credits Received | Fee | Bonus |
|---|---|---|---|---|
| Starter | $10 | $9.45 | $0.55 (5.5%) | -- |
| Standard | $50 | $47.25 | $2.75 (5.5%) | -- |
| Professional | $100 | $94.50 | $5.50 (5.5%) | BYOK limit increased to 2M |
| Enterprise | Custom | Custom | Negotiated | Dedicated support, SLA |
Per-token prices in USD. Data fetched from API in real-time.
| Model | Provider | Input ($/1M tokens) | Output ($/1M tokens) |
|---|---|---|---|
| Loading... | |||
Standard mode: You pay provider_cost x (1 + 5.5%). The 5.5% covers payment processing and gateway fees.
BYOK mode: You pay nothing for the first 1M requests per month. After that, you pay provider_cost x 5% as a routing fee.
In both cases, you pay only for what you use. No monthly subscription fees.
Yes. You can switch between modes per request by including or omitting the provider_key field. Both modes are available on the same account.
We accept major credit and debit cards via Stripe. Cryptocurrency (USDC) support is coming in Q3 2026.
Credits expire 365 days from your last account activity. Inactive accounts are notified via email before expiration.
Every API request counts as one request regardless of token count. The free tier (1M/month) is shared across all models. The counter resets on the 1st of each month at 00:00 UTC.
Unused credits are refundable within 14 days of purchase. Please contact our support team to initiate a refund.