Qwen is the Alibaba model family. Some versions are open, which makes Qwen one of the most popular models with developers.
The NeuralBox chat has Qwen 3.8 Max, 3.7 Plus and 3.8 Flash. Qwen Image and Wan video from the same company have their own pages.
Available versions
Qwen3.7 Plus
per exchange ≈ $0.004
Balanced price and quality. Context 1M tokens.
Qwen3.8 Max (0902)
per exchange ≈ $0.020
The flagship. Context 1M tokens.
Qwen3.8 Flash
per exchange ≈ $0.002
Fast and cheap. Context 1M tokens.
What it does
Code
Writes and explains code.
Long context
Up to 1M tokens.
Cheap
Flash costs a fraction of a cent per answer.
Prompt examples
Attach files and images to any message.
Table
Make a monthly family budget table with categories
Code
Write a Python Telegram echo bot
Parameters
Billing
per answer, no subscription
Attachments
files and images in a message
Switching
change the model any time
Context
up to 1M tokens
How much Qwen costs
Model
Exchange, tokens
USD
Qwen3.7 Plus
122
≈ $0.004
Qwen3.8 Max (0902)
608
≈ $0.020
Qwen3.8 Flash
47
≈ $0.002
One exchange here means a 1,000-token question and a 1,000-token answer, about 700 words together. USD at the Basic rate: $5 = 150,000 tokens. Long threads and model reasoning cost more. The exact price of each answer shows in the chat.
$5 gives 150,000 tokens. With Qwen3.8 Flash that is about 3,191 questions with a full answer.
How to start
Sign in
With Google, Telegram or email. Takes under a minute.
Top up
From $5 by card or from $2 in crypto. No subscription, and tokens never expire.
Generate
Open the Text tab and pick Qwen. Ask a question or attach a file.