Alibaba ● Coming soon

Qwen3.8-Max

qwen3.8-max

Alibaba's largest Qwen flagship yet: 2.4T parameters, up to 1M tokens of context

Qwen3.8-Max is the Qwen flagship Alibaba released on August 3, 2026. It is a sparse mixture-of-experts model with 2.4 trillion parameters, about 95 billion of them active per token, a context window of up to 1 million tokens and image input, built for autonomous coding, long-horizon tasks and everyday knowledge work. It is also the first Max-class Qwen model whose weights Alibaba has opened. NezhaGate is integrating it now: leave your email and we will tell you the day it goes live.

Released by the vendor: 2026-08-03 · not yet callable on NezhaGate

Highlights

2.4T-parameter sparse MoE with about 95B parameters active per token
Up to 1M tokens of context: a whole codebase or a long report in one request
Multimodal: text and image input, and Alibaba also cites video understanding
Built for agentic coding and long-horizon tasks, plus research and office work
The first Max-class Qwen model released with open weights

Use cases

Reading, refactoring and fixing large codebases

Agents that plan in several steps and call tools repeatedly

Analysis and Q&A over long contracts, filings and papers

Making sense of mixed text-and-image material

Bilingual Chinese and English writing and content production

FAQ

When will Qwen3.8-Max be available on NezhaGate?

We are integrating and testing it now. The moment it is callable this page switches to live and shows the price. Leave your email and we will write that day.

How will I call it?

Like every other model here, through the OpenAI-compatible API: set base_url to https://nezhagate.com/v1 and model to qwen3.8-max. Your existing OpenAI SDK code works unchanged.

Is there a Qwen model I can use today?

Yes. Qwen3.7 Max (qwen3.7-max) is live on NezhaGate and callable right now.