Qwen3.5 397B A17B
Neural Network
Qwen3.5 397B A17B is Alibaba's native vision-language model for text tasks and image analysis via a unified Qwen API.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Qwen3.5 397B A17B?
Qwen3.5 397B A17B is a native vision-language model from Alibaba's Qwen3.5 series, designed to handle both text and images natively. Its hybrid architecture combines a linear attention mechanism with a sparse mixture of experts: out of 397 billion parameters, about 17 billion are activated per step, making inference more computationally efficient. In BotHub, it is available immediately without VPN or foreign cards, with payment in rubles via Russian cards and pay-as-you-go billing for non-expiring tokens. Access over 250 neural networks in one window to easily run and compare the same prompt across multiple models, while a unified OpenAI-compatible API allows switching between them without rewriting integrations. Data is transmitted with AES-GCM encryption, and companies have access to contracts, invoices, electronic document management, and an admin panel with limits. It is useful for analyzing interface screenshots or diagrams, converting photos of documents and tables into structured data, debugging code from logs, building multimodal services via API, and handling repetitive tasks.Frequently asked questions about Qwen3.5 397B A17B
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
This is a Qwen text model with a 256,000-token context: it writes and refactors code, analyzes errors, solves math and logic problems, and processes long documents and input images. Suitable for analytics, codebase work, and agentic scenarios with function calling.
The single response limit is 235,929 tokens, within a total context of 256,000. This is enough to output a large code module, a detailed analysis, or an entire voluminous document without breaking the task into parts. Remember: long output consumes more tokens, which are charged on a pay-as-you-go basis.
Yes, reasoning before answering is confirmed: the model first builds a chain of steps and then formulates the conclusion. This significantly helps in math, logic, code debugging, and tasks where not only the result but also the correct path to it is important.
Yes. The model supports function calling and structured JSON output, so it can be placed in agentic pipelines and connected to external services. In BotHub, it is available via a unified OpenAI-compatible API: you won't have to rewrite integration schemes when switching models.
According to our data, images and documents are accepted as input: you can send an interface screenshot, diagram, chart, or PDF and ask to analyze the content. Audio and video support is not noted, so for such files, it is better to choose a specialized model in the same window.
We do not separately track the model's language characteristics, so we cannot provide exact ratings — test it with your own task: the BotHub interface is in Russian, access works from the Russian Federation without a VPN or foreign card, and payment is accepted via Russian cards. If the result is not satisfactory, switch to another model in the same window.