Pay-As-You-Go Pricing

Only pay for what you use. No monthly commitments, no hidden fees.

All prices in CNY (Chinese Yuan)

Start Building Today

Add credit to your account and only pay for what you use

Current Balance

¥—

New users get 5,000 tokens free to get started

Model Pricing & Performance

Per-token billing with no monthly minimum. GTC-2.5 mini and TFMF are live today; the companion models are already priced and unlock once LoRA training completes.

GTC-2.5 mini

Live

Lightweight 64M base model for basic conversation, knowledge Q&A and simple text generation. Low cost, low latency and built for high-frequency calls.

Input Tokens

¥0.5 per 1M tokens

Text you send to the model

Output Tokens

¥1 per 1M tokens

Text generated by the model

Specifications

Parameters: 64M
Speed: Very Fast
Best for: Lightweight apps & edge

TFMF

Live

TAI Research's language model, post-trained on Qwen3.5.

Input Tokens

¥0.5 per 1M tokens

Text you send to the model

Output Tokens

¥3 per 1M tokens

Text generated by the model

Specifications

Base: Qwen3.5, post-trained
Adaptation: LoRA supported
Best for: General & vertical domains

Companion Models

Purpose-built companions fine-tuned on TFMF with LoRA. Currently in training — pricing is already fixed at ¥1 input / ¥5 output per 1M tokens.

Giggle

In training

General companion for easy small talk, playful humour and everyday listening — a light-hearted style made with younger users in mind.

Base: TFMF + LoRA
Input: ¥1 / 1M tokens
Output: ¥5 / 1M tokens

Bro

In training

Companion for teenage boys — buddy-style conversation about gaming, sports, study and tech, with a direct and down-to-earth tone.

Base: TFMF + LoRA
Input: ¥1 / 1M tokens
Output: ¥5 / 1M tokens

Sweetie

In training

Companion for teenage girls — gentle conversation for emotional support, relationships and everyday sharing. Considerate, empathetic and a good listener.

Base: TFMF + LoRA
Input: ¥1 / 1M tokens
Output: ¥5 / 1M tokens

Token Calculator

Estimate Your Costs

See how much your usage might cost based on your expected token consumption.

Estimated Cost

¥0.00

Based on your inputs and selected model

Frequently Asked Questions

How does pay-as-you-go billing work?

You add credit to your account, and we deduct costs based on your actual usage. There are no monthly subscriptions or commitments. You only pay for the tokens you consume.

What's the difference between input and output tokens?

Input tokens are the text you send to our models. Output tokens are the text generated by our models. Output tokens typically cost more because generating text requires more computational resources.

Is there a free tier?

Yes! All new accounts receive 5,000 tokens completely free to test our API and build your first integration.

What payment methods do you accept?

We accept Alipay, WeChat Pay, and major credit cards. All transactions are processed securely.

How accurate is the token calculator?

The calculator provides estimates based on standard tokenization. Actual usage may vary slightly depending on your specific inputs and model parameters.

Built by Students, For Developers

We're a team of high school students passionate about AI. Despite limited resources, we've created TAI to make powerful AI models accessible to everyone. Your support helps us continue improving and expanding our services.