Qwen4-Max is a hosted preview with no weights, shutting off Nov 25 — local alternatives mapped by RAM
alex_verem · x · 2026-11-17
Alibaba's Qwen4-Max has launched as a hosted-only preview: no downloadable weights, and the preview shuts off on November 25. (Unverified specs — await official confirmation.)
The post pairs this with a local-deployment cheat sheet using free, permanent open-weight Qwen models:
- 16GB laptop → Qwen3.5 9B for chat and writing
- 32GB laptop → Qwen3.8 27B for coding and harder tasks
- Gaming PC → Qwen3.8 Flash Next for speed
- Company server → Qwen3.8 Max, the 2.4T-parameter flagship
Takeaway: the hosted preview is temporary; pick an open model that fits your machine's memory instead.
More from Infra
- Over 45 Companies Including Major Names Hit by Simultaneous Outages — Serasul · 2026-11-17
- Indie dev's cloud bill for Pearl Chat is so steep he jokes about a poultry farm — weskambale · 2026-11-17
- Open-Source Tokspur Engine Claims 4x Faster Mac Inference Than Ollama.cpp — SimonSanDigital · 2026-11-17
- ExLlamaV3 MoE offloading on 9070XT doubles prefill to 2606 t/s for Qwen3.6 35B — willeyh · 2026-11-17
- BIS 132-page export control rule closes cloud-rental, smuggling and TSMC loopholes — ChinaTalk · 2026-11-17
- Engineer runs Gemma 3 270M entirely inside an office laser printer at 60s per token — osanseviero · 2026-11-17