Qwen4-Max Preview: Alibaba's next flagship, open for 14 days only
Alibaba opened a limited preview of Qwen4-Max, the first model of its Qwen4 family, with a 1M-token context window and a 393K-token output ceiling, available through Token Plan and Qwen Chat until it retires on 25 November 2026.
Why it matters
The first Qwen4 flagship anyone can actually try: 1M context, text/image/video input, and a hard 14-day expiry that makes it a pure evaluation window, not a dependency.
What you could build with it
A startup evaluating long-context models could run a two-week benchmark sprint against Qwen4-Max Preview: feed it full codebases and hour-long meeting recordings, score retrieval and reasoning against the current production model, and decide before the 25 November retirement whether the Qwen4 line earns a place in the routing table.
Does it hold up?
Too early to judge: the preview opened this morning, no benchmarks exist yet, and access ends 25 November — the only honest verdict now is to run a fixed evaluation set against it before it retires.
Built with Qwen4-Max Preview
- Qwen4-Max Preview in Qwen Chat: a second door, same 14 daysarticle · Deep dive on the two access routes (Token Plan vs Qwen Chat), the 1M context and 393K output ceilings, the reasoning_effort/max_tokens defaults that break ported requests, and the 25 November retirement.
- Alibaba's Qwen4-Max Preview Offers a 1M-Token Context Window, but Access Ends Nov. 25article · Specs and limits roundup: five access points (Token Plan, QwenWork, Qoder, Qwen app, QwenStudio), no published parameter count or benchmarks, and the interactive-use-only licence.
- Alibaba opens Qwen4-Max preview with a November 25 cutoffarticle · Coverage of the fixed evaluation window, the Token Plan subscription terms, and why the preview is an evaluation target with an expiry date rather than a production dependency.
Learn more
First spotted on article: source.
More AI models
VoiceStudio: open-source, fully-local ElevenLabs alternative for voice cloning, dubbing, and audiobooksdebpalash/VoiceStudio is a desktop app that runs voice cloning, voice design, video dubbing, dictation, transcription…models · JEV 0.75Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Google AI Edge Gallery: offline speech recognition and synthesis on-deviceGoogle added fully offline speech recognition and text-to-speech to AI Edge Gallery, with ML Kit APIs on Android and…models · JEV 0.73Soup: fine-tune an 8B LLM on a 4 GB laptop GPU from one YAML fileAn open-source CLI that fine-tunes LLMs from a single YAML file, with layer streaming that trains an 8B model on a 4…models · JEV 0.73Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.