Qwen 4 Release Date: Officially “in Training”, No Date Announced
As of 2026-11-05, Alibaba has not announced a release date for Qwen 4; its official statement is the 2026-09-22 press release saying the next-generation model Qwen 4 is currently in training. The newest series in service is Qwen3.8.
Updated 2026-11-05
Claude, GPT and Chinese models on one key, billed per token; check /models for callable models.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Four key points
What Alibaba says about Qwen 4
Alibaba's press release says Qwen 4 is currently in training and gives no release date for it.
Date of Alibaba's press release
A press release issued around the Apsara Conference; the same release mentions a roadmap for the later Qwen 4.5 and Qwen 5 series, also without dates.
Qwen 4 on official model pages
On 2026-11-05 no model named Qwen 4 appears in Qwen's GitHub repository, its Hugging Face organization page or the Alibaba Cloud Model Studio model overview.
Latest series per the official repo
Qwen's GitHub repository calls Qwen3.8 the latest series; open weights for Qwen3.8-2.4T-A95B and Qwen3.8-27B were posted on 2026-08-12 and 08-14.
What this question is asking
As of 2026-11-05, Alibaba has not announced a Qwen 4 release date. The official statement is the 2026-09-22 press release: at the conference Alibaba revealed that its next-generation model Qwen 4 is currently in training, and set out a roadmap for the later Qwen 4.5 and Qwen 5 series. The newest series in service is Qwen3.8; the Alibaba Cloud Model Studio model overview (page updated 2026-11-02) lists text models such as qwen3.8-max, qwen3.7-plus and qwen3.8-flash, and no Qwen 4.
What Alibaba has said officially
In its press release from Hangzhou on 2026-09-22, Alibaba wrote that its next-generation model, Qwen 4, “is currently in training”. The release gives no release date for Qwen 4 and says nothing about whether it will be open-sourced.
Official timeline
Qwen's GitHub update log: open weights for Qwen3.8-2.4T-A95B became available on Hugging Face and ModelScope.
Same update log: Qwen3.8-27B became available on Hugging Face and ModelScope; it is the latest model entry in that log.
Alibaba's press release disclosed that Qwen 4 is currently in training and laid out a roadmap for the Qwen 4.5 and Qwen 5 series, without dates.
Officially confirmed vs community speculation
Officially confirmed
All of this can be checked in the official pages fetched on 2026-11-05: Alibaba Cloud's press release (2026-09-22) says Qwen 4 is currently in training and sets out a Qwen 4.5 and Qwen 5 roadmap, with no date and no mention of open-sourcing; Qwen's GitHub repository calls Qwen3.8 the latest series, and its update log has no Qwen 4; the Alibaba Cloud Model Studio model overview (page updated 2026-11-02) has no Qwen 4 either. On the official pages Qwen 4 is “in training”, not “released”.
⚠️ Community speculation (not confirmed by Alibaba)
The community has all kinds of speculation about Qwen 4's timing and specs (for example Emergent, 2026-10-30); Alibaba has confirmed none of it.
On official pages vs not
Has official pages (in service)
Models in service: the hosted Qwen3.8 and Qwen3.7 models in the Model Studio model overview, plus Qwen3.8-27B and Qwen3.8-2.4T-A95B with open weights.
No official page
Qwen 4: its official status is “in training”; there is no model ID, open weights or release date yet.
What to do now
① To check whether Qwen 4 has shipped, look at three places: Alibaba Cloud's press releases, the update log in Qwen's GitHub repository and the Model Studio model overview; once Alibaba publishes it, this page will be updated with the official date and model ID. ② To use Qwen today: the model overview describes qwen3.8-max as for complex reasoning and coding, qwen3.7-plus as balanced across performance, speed and cost, and qwen3.8-flash as fast and cost-effective. ③ Through QCode, Anthropic-protocol clients set ANTHROPIC_BASE_URL to https://api.qcode.cc/api and OpenAI Chat clients set base_url to https://api.qcode.cc/openai/v1. ④ Do not put an ID such as qwen4 into a production config: no official doc lists it.
On QCode
Getting started in three steps: ① create an account; ② top up (Alipay, WeChat Pay, cards, crypto); ③ follow the docs.qcode.cc endpoints page: https://api.qcode.cc/api for the Anthropic protocol, https://api.qcode.cc/openai/v1 for OpenAI Chat, your QCode key, and a model such as qwen3.8-max. As of 2026-11-05, QCode's /models lists qwen3.8-flash, qwen3.8-max and qwen3.7-plus, and no Qwen 4; billed per token with unit prices on /models. Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days; if you already have a Personal Plan, add a booster pack when you need more quota.
Frequently asked questions
Is Qwen 4 out?
No. As of 2026-11-05 Alibaba has not released Qwen 4: the official statement is “currently in training” in the 2026-09-22 press release, and neither the GitHub update log nor the Model Studio model overview lists it.
When is the Qwen 4 release date?
As of 2026-11-05 Alibaba has not announced one. There is community speculation (for example Emergent, 2026-10-30), which Alibaba has not confirmed; such claims are not a release date.
Will Qwen 4 be open source?
Alibaba has not said. The 2026-09-22 press release covers only that Qwen 4 is in training and the later roadmap, and does not mention open-sourcing or a license. The released Qwen3.8 does have open weights: the GitHub repository says Qwen3.8 brings a Qwen-Max-class model to open release for the first time, and Qwen3.8-27B and Qwen3.8-2.4T-A95B are available on Hugging Face and ModelScope.
Can I use Qwen3.8 and Qwen3.7 now?
Yes. The Model Studio model overview (page updated 2026-11-02) still lists qwen3.8-max, qwen3.7-plus and qwen3.8-flash, and QCode's /models has all three on sale.
How do I call Qwen on QCode?
Use your QCode key and set model to qwen3.8-max, qwen3.8-flash or qwen3.7-plus. docs.qcode.cc says Chinese models such as Qwen work on both the Anthropic endpoint and the OpenAI Chat endpoint: the base URL is https://api.qcode.cc/api for the former and https://api.qcode.cc/openai/v1 for the latter.
How do I pay from China?
QCode accepts Alipay, WeChat Pay, cards, Apple Pay, Google Pay and crypto; all but crypto usually arrive within a minute, while crypto is checked by hand and credited within one day. The full set-up steps for China are on /claude-code-china-guide.
Sources
Alibaba Cloud's press release of 2026-09-22, Qwen's GitHub repository (QwenLM/Qwen3.8), the Qwen organization page on Hugging Face and the Alibaba Cloud Model Studio model overview were fetched and archived on 2026-11-05; the community-speculation part cites Emergent (2026-10-30), not confirmed by Alibaba. The QCode side comes from qcode.cc /models, the Pay-as-you-go page, the top-up page copy and the docs.qcode.cc endpoints page (2026-11-05).
Use Qwen3.8 today
One QCode key switches between qwen3.8-max, qwen3.8-flash and qwen3.7-plus by changing model, billed per token with unit prices on /models.
Related reading
Pay-as-you-go Packs
Buy a one-time quota instead of a monthly plan and use it within 7 days; per-model rates are on /models.
Qwen3.8-Max guide
Alibaba's Qwen3.8-Max: model ID, context length and how it is listed in the QCode catalogue.
2026 AI Model Radar
A model-by-model tracker of what is available and what is still in preview or restricted, updated as vendors publish.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Checked on 2026-11-05. Models and specs follow Alibaba's official pages; QCode availability and unit prices follow /models. This page will be updated once Alibaba publishes.