Alibaba launches Qwen4-Max preview with 1 million-token context

Alibaba announced on 11 Nov that its next-generation Qwen4 models are arriving and opened a preview of Qwen4-Max-Preview, available only through Alibaba’s Token Plan until 25 Nov 2026 IT Home Report. The preview offers a 1 million-token context window and caps output at 393,216 tokens, shared between reasoning and the final answer.

Alibaba Cloud’s Model Studio documentation describes Qwen4-Max-Preview as the flagship multimodal model in the Qwen family. It combines text generation, image and video understanding, deep reasoning, coding assistance, office work, front-end design and complex agentic tasks. Thinking is on by default and can be adjusted with a reasoning-effort setting. The model supports web search, function calling, built-in tools and structured JSON outputs, and is compatible with OpenAI’s Chat Completions and Responses APIs and Anthropic’s Messages API.

Access is narrow by design. The model is reachable only via the Token Plan subscription, not pay-as-you-go. The documentation sets the preview from 10:00 on 11 Nov to midnight at the end of 25 Nov, Beijing time, and states the model will not change during that window and will be retired afterward. For the preview, Alibaba says Qwen4-Max calls consume Token Plan credits at a promotional rate as low as 20 % of the standard rate.

Alibaba has not released benchmark results, a model card, a parameter count, a per-token price or weights for Qwen4-Max, and has not dated the full Qwen4 launch. The capability claims are the company’s own.

Our read: The preview gives enterprises a short-term sandbox to probe ultra-long context and multimodal features, but the absence of performance data and the subscription-only access constrain its relevance for production workloads.