Mercury diffusion LLM API, 100M free tokens for new accounts
About this product
Inception is a US AI company whose founding team comes from Stanford, UCLA and Cornell and helped invent key techniques behind modern AI, including diffusion models, Flash Attention and DPO. Its Mercury models are diffusion large language models (dLLMs): instead of writing one token at a time like ChatGPT, they generate many tokens in parallel, with time to first token under 0.3 seconds, 5–7x higher throughput and lower cost per task.
The latest Mercury 3 offers a 1M-token context window, adjustable reasoning effort (instant / low / medium / high), tool calling and structured outputs. The family also includes Mercury Voice for voice agents, Mercury Edit 2 for code completion and edits, and Mercury Decide for classification. The API is OpenAI-compatible, so you can swap it into an existing app by changing the base URL and model name.
Which option to choose
Account: sign in to the Inception Platform (API dashboard) and create your own API keys. New accounts get 100 million free tokens from Inception, and the free tier allows 1,000 requests per minute (subject to Inception's current policy).
US$100: adds US$100 of usage to the account, charged by model and token usage.
What you can do with Mercury
Chat · Beginner
Instant chat and reasoning
Under 0.3 s to the first token — great for support bots and assistants that must reply instantly.
Create an API key in the dashboard
Call the mercury-3 model and set the reasoning effort
Show the reply in your chat interface
Try asking
Explain diffusion language models in three sentences, then give an everyday example.
Integrate · Beginner
Swap into an existing app
OpenAI-compatible — just change the base URL and model name.
Set the base URL to https://api.inceptionlabs.ai/v1
Change the model name to mercury-3
Use your Inception API key and run
Coding · Intermediate
Real-time code completion
Mercury Edit 2 handles completion and edits fast enough to keep you in flow.
Collect the code before and after the cursor
Call the fill-in-the-middle or edit endpoint
Insert the returned code into the editor
Agents · Advanced
Voice and agent workflows
Low latency plus tool calling for voice agents and multi-step automation.
Define the tools (functions) in the request
Let the model choose a tool and return its arguments
Run the tool and send the result back to finish the task
Getting started
Create a key: after signing in, go to API Keys, create a new key and keep it safe.
Call the API: base URL https://api.inceptionlabs.ai/v1, model mercury-3; works with the official OpenAI SDK or any OpenAI-compatible tool.
Use your existing tools: LiteLLM, LangChain, the Vercel AI SDK and other tools with a custom OpenAI endpoint just need the base URL and API key.
Code completion: use the Mercury Edit 2 fill-in-the-middle (FIM) and edit endpoints for real-time completion in an editor.
Tips
Set reasoning effort to instant or low for quick answers; use medium or high for complex tasks.
Check usage and remaining credit in the dashboard, and top up only when the free tokens run low.
[Inception homepage]
[Mercury models and prices]
[Docs: 100M free tokens for new accounts]
Check before buying
Charged by token usage; top-ups are non-refundable. Free tokens are subject to Inception's current policy.
🙋 Manual delivery⏰ Working hours UTC+8 09:00–21:00📧 Sent to your order email, also viewable once on the order page
What AI services does 5buyai.com support? Can I get an invoice?
Accounts, subscriptions and top-ups for ChatGPT, Claude, Midjourney, Gemini, Grok and other major AI tools. Invoices are available for AI orders paid in Chinese yuan (CNY).
How do I receive my order after paying?
Orders are delivered manually: after successful payment, the account or card code is sent to your order email during working hours (9:00–21:00, UTC+8). You can check your order details anytime on the order lookup page by order number, email or browser cache.
What payment methods are supported?
Alipay, WeChat Pay, USDT, and Credit Card payments. Credit card payments require first topping up your balance, and then using the balance to purchase products.
Do I need an account to buy?
No. You can check out as a guest and receive items by email; registering lets you manage orders, balance and invoices in one place.
Are prices shown in CNY?
All prices are listed in Chinese Yuan (¥). When paying with USDT the amount is converted automatically at the live rate.
Is there after-sales support?
After-sales support is available for some products, handled via live chat or the order dispute page. Accounts are virtual products: once delivered, no refunds are given unless there is an after-sales issue. CNY refunds and USD refunds are only made back to the original payment method, and USDT refunds are only made on the BSC network.
I am a business buyer and need to make a bulk purchase. How can I get in touch?
Please click the chat widget at the bottom-right corner to reach us and leave your email or other contact details. We will get back to you as soon as possible.
No reviews yet — be the first to review after your purchase.
Disclaimer
This is an independent third-party platform, not affiliated with, endorsed by, or authorized by any brand mentioned. We do not host any third-party login page and do not collect third-party credentials.
