AI model · Alibaba Qwen
Qwen 4 Max Preview, Explained: Specs, Access, Price
Qwen 4 Max is Alibaba's next flagship Qwen model, available since November 11, 2026 only as a time-limited preview through Alibaba's Token Plan subscription. Synexa does not serve it; use Synexa when your app also needs images, video, audio or 3D from one API key.
- 1M-token context window
- Preview ends November 25, 2026
- Synexa: image, video, audio, 3D APIs
What is Qwen 4 Max?
Qwen 4 Max, written Qwen4-Max by Alibaba, is the top model of the coming Qwen 4 generation of large language models. On November 11, 2026 the Qwen team posted "Qwen4 is coming" and opened a preview of Qwen4-Max at the same time. Alibaba promises deeper knowledge, richer world understanding and stronger coding and agentic skills, aimed at real-world tasks from complex coding and knowledge work to front-end builds and professional software workflows.
What you can use today is a preview, not a release. The preview runs on Alibaba's Token Plan subscription with the model ID qwen4-max-preview, and Alibaba's own apps got it early. Alibaba has not published a parameter count, benchmark scores, a per-token API price or a date for the full launch, and no Qwen 4 weights are on Qwen's Hugging Face page as of November 13, 2026.
Qwen 4 Max preview specs: context, output, inputs
Everything below comes from Alibaba Cloud's model documentation as of November 13, 2026.
- Maker
- Alibaba (Qwen team)
- Status
- Preview, November 11 to 25, 2026
- Model ID
qwen4-max-preview(aliasqwen4.0-max-preview)- Capabilities
- Reasoning, visual understanding, text generation
- Input
- Text, images and video
- Output
- Text
- Context window
- 1M tokens
- Max output
- 393,216 tokens, thinking and answer combined
- Images
- Up to 16M pixels each, up to 2,048 images
- Video
- Up to 2 hours and 2 GB per video, up to 64 videos
- Features
- Thinking mode, function calling, built-in tools, structured output
- Parameters / weights
- Not published
How does the Qwen4 preview differ from Qwen3.8-Max?
Alibaba's docs list the new preview next to its current models. On paper the main change is a much longer answer budget.
Context window
- qwen4-max-preview
- 1M
- qwen3.8-max
- 1M
- qwen3.7-plus
- 1M
Max output
- qwen4-max-preview
- 393,216
- qwen3.8-max
- 128K
- qwen3.7-plus
- 128K
Image and video input
- qwen4-max-preview
- Yes
- qwen3.8-max
- Yes
- qwen3.7-plus
- Yes
Thinking, tools, structured output
- qwen4-max-preview
- Yes
- qwen3.8-max
- Yes
- qwen3.7-plus
- Yes
Alibaba's selection guide
- qwen4-max-preview
- Quality first
- qwen3.8-max
- not in the guide
- qwen3.7-plus
- Balanced choice
| Spec | qwen4-max-preview | qwen3.8-max | qwen3.7-plus |
|---|---|---|---|
| Context window | 1M | 1M | 1M |
| Max output | 393,216 | 128K | 128K |
| Image and video input | Yes | Yes | Yes |
| Thinking, tools, structured output | Yes | Yes | Yes |
| Alibaba's selection guide | Quality first | not in the guide | Balanced choice |
Alibaba's selection guide recommends the preview when quality comes first (Token Plan only), qwen3.7-plus for balance and qwen3.8-flash for cost.
How much does the Qwen 4 Max preview cost?
There is no per-token price. You pay for a Token Plan subscription, and each call uses Credits from its monthly quota. International prices from QwenCloud, as of November 2026.
Lite
- Monthly price
- $6
- New-user price
- $3
- Credits per month
- 20,000
Essential
- Monthly price
- $10
- New-user price
- $5
- Credits per month
- 45,000
Standard
- Monthly price
- $18
- New-user price
- not listed
- Credits per month
- 80,000
Pro
- Monthly price
- $43
- New-user price
- not listed
- Credits per month
- 200,000
Credit pack (add-on)
- Monthly price
- $15
- New-user price
- not listed
- Credits per month
- 20,000
| Token Plan (Individual) | Monthly price | New-user price | Credits per month |
|---|---|---|---|
| Lite | $6 | $3 | 20,000 |
| Essential | $10 | $5 | 45,000 |
| Standard | $18 | not listed | 80,000 |
| Pro | $43 | not listed | 200,000 |
| Credit pack (add-on) | $15 | not listed | 20,000 |
During the preview, Qwen4-Max calls use as little as 20% of the standard Credits rate, and 8% between 22:00 and 08:00 (UTC+8). Credits per call depend on tokens, thinking and tool use, and Alibaba lists no fixed rate. New-user prices run to November 30, 2026; Individual plans can't be cancelled. In mainland China the same tiers cost ¥39, ¥79, ¥139 and ¥312.
How to try Qwen 4 Max today
Alibaba names two routes: the Token Plan for developers, and its own apps for everyone else.
Subscribe to Token Plan
Use Alibaba's apps
Plan for the cutoff
Is there an official Qwen 4 Max benchmark?
No. As of November 13, 2026 Alibaba has published no Qwen 4 Max benchmark scores. The launch post and the Token Plan docs describe goals (deeper knowledge, world perception, coding, agents) and early results on real tasks, without numbers. Scores you see on YouTube, leaderboards or reseller pages come from third-party tests of a preview model that will be retired in two weeks, so treat them as anecdotes.
The nearest official data point is architectural. In August 2026 the Qwen team released Qwen3.8-Flash and called it an early preview of the Qwen4 architecture, with hybrid GDN and QSA attention, gated residuals, n-gram embeddings and the Muon optimizer. Alibaba has not said whether the Qwen4-Max preview uses that exact design. If you need numbers for a decision, run your own evals on the preview before November 25.
What to know before you build on the Qwen4 preview
The preview is useful for testing, with a few limits written into Alibaba's rules.
- Test real tasks now; the preview won't change until it ends
- Schedule long sessions in the night window (22:00 to 08:00 UTC+8) at 8% of standard Credits
- Check per-call Credits in the console's model usage view
- Keep a fallback such as qwen3.8-max wired in
- Ship production features on the preview ID; it goes offline after November 25
- Expect a pay-as-you-go API price; Alibaba lists the preview as Token Plan only
- Quote benchmark numbers as official; none are published
- Call it from an app backend or batch script; Individual plan terms allow interactive use in coding and agent tools only
Can you run Qwen 4 Max on Synexa?
No. Synexa does not serve Qwen4-Max or any other Qwen language model, and it offers no chat LLMs; its only text outputs are LLaVA-13B image answers, Whisper speech transcripts and an NSFW image-check label. For the language side, use Alibaba's Token Plan during the preview, inside a supported coding or agent tool.
Synexa covers the media half of an AI product through one REST API. One key gives you image models such as Nano Banana Pro at $0.10 a request, text-heavy images with GPT Image 2 at $0.10, plus video, speech, upscaling and image-to-3D models such as Hunyuan3D 3.1 at $0.16 a request. A common pattern: let the language model plan and write the prompt, then send it to Synexa to render the asset. Prices are on each model page in Explore, new accounts start with $0.10 of free credit, and the quickstart shows the first call.
- Pay per request, no subscription
- Python and Node clients, plus plain HTTP
- Image, video, audio, upscaling and 3D in one catalog
Synexa models to pair with a Qwen 4 Max agent
Flat per-request prices as listed on Synexa model pages, November 2026.
Product or marketing image from a written brief
- Synexa model
- Nano Banana Pro
- Price
- $0.10 / request
Image or edit with sharp, readable text
- Synexa model
- GPT Image 2
- Price
- $0.10 / request
30-second music clip from a text prompt
- Synexa model
- Lyria 2
- Price
- $0.10 / request
Textured 3D model from one image
- Synexa model
- Hunyuan3D 3.1
- Price
- $0.16 / request
Fast draft 3D mesh
- Synexa model
- TRELLIS
- Price
- $0.02 / request
Upscale or restore a photo
- Synexa model
- Real-ESRGAN
- Price
- $0.0015 / request
Transcribe a voice note
- Synexa model
- Whisper v3
- Price
- $0.004 / request
| Job your agent hands off | Synexa model | Price |
|---|---|---|
| Product or marketing image from a written brief | Nano Banana Pro | $0.10 / request |
| Image or edit with sharp, readable text | GPT Image 2 | $0.10 / request |
| 30-second music clip from a text prompt | Lyria 2 | $0.10 / request |
| Textured 3D model from one image | Hunyuan3D 3.1 | $0.16 / request |
| Fast draft 3D mesh | TRELLIS | $0.02 / request |
| Upscale or restore a photo | Real-ESRGAN | $0.0015 / request |
| Transcribe a voice note | Whisper v3 | $0.004 / request |
Synexa bills each request from a prepaid balance with no monthly plan; after the free credit, top-ups start at $100.
Should you switch to Qwen 4 Max now?
Try it, don't switch. The Qwen 4 preview is a two-week test with no published benchmarks, no per-token price and a fixed shutdown date, so it is a good time to run your own task suite and a poor time to migrate production. The Lite plan at $6 a month, or $3 for new users, is a cheap way to get access, and preview calls burn Credits at a fifth of the normal rate.
For production work today, Alibaba's selection guide points to qwen3.7-plus for balance and qwen3.8-flash for cost. Watch for the full Qwen 4 release, which Alibaba says is coming but has not dated. If your app also produces images, music or 3D assets, see how a dedicated media API compares in our WaveSpeed AI alternative page, read about another new model, Claude Fable 5.2, and check per-request prices on pricing.
Related pages
More model explainers and API comparisons.
Frequently asked questions
What is Qwen 4 Max?
Qwen 4 Max, or Qwen4-Max, is Alibaba's next flagship Qwen language model and the top tier of the coming Qwen 4 generation. Since November 11, 2026 it has been available as a preview with the model ID qwen4-max-preview. It reads text, images and video, writes text, and has a 1M-token context window with up to 393,216 output tokens.
Is Qwen 4 Max released?
Not fully. Alibaba opened a Qwen4-Max preview on November 11, 2026 and said Qwen4 is coming, but gave no date for the full release. The preview runs until 24:00 on November 25, 2026 (UTC+8) and is then taken offline. Alibaba has not published parameters, benchmark scores or open weights for any Qwen 4 model.
How do I get access to the Qwen 4 Max preview?
Subscribe to any Token Plan Individual tier on QwenCloud, or an Individual or Team plan on Alibaba Cloud Model Studio in mainland China, then call the preview model ID with your plan's base URL and API key. Alibaba also lists early access in QwenWork, Qoder, the Qwen app and QwenStudio. Alibaba's docs mark the model as Token Plan only, so it isn't on the pay-as-you-go API.
How much does Qwen4-Max cost?
Alibaba publishes no per-token price for the preview. Access comes with a Token Plan subscription: $6 a month for Lite with 20,000 Credits, $10 for Essential, $18 for Standard and $43 for Pro, with new-user prices of $3 and $5 until November 30, 2026. During the preview, Qwen4-Max calls use as little as 20% of the standard Credits rate.
Does Qwen 4 Max have benchmark scores?
Not official ones. As of November 13, 2026 Alibaba has released no benchmark numbers for the Qwen4-Max preview. Its announcement describes stronger knowledge, world understanding, coding and agentic skills and early results on real tasks, without scores. Numbers on third-party sites come from independent tests of a short-lived preview, so run your own evals if you need a basis for a decision.
Can I download Qwen4 weights?
No. As of November 13, 2026 there are no Qwen 4 models on Qwen's Hugging Face page, and the preview is cloud-only through Token Plan. The closest open release is Qwen3.8-Flash-Next from August 2026, which the Qwen team described as an early look at the architecture it is exploring for Qwen4. Alibaba has not said whether open weights will follow.
What is the difference between qwen4-max-preview and qwen4.0-max-preview?
Nothing. Alibaba's docs list qwen4.0-max-preview as an alias of qwen4-max-preview. Both model IDs work in Token Plan requests, and bills show the usage as qwen4.0-max-preview either way. Whichever you use, the preview model is retired after November 25, 2026, so keep the ID in one config value that is easy to change.
Can I use Qwen 4 Max through Synexa?
No. Synexa does not serve any Qwen language model or any other chat model. Synexa is a pay-per-request API for generative media: images, video, music and speech, upscaling and image-to-3D. Many teams call a language model for planning and writing, then call Synexa to render the image, clip, track or 3D model. New Synexa accounts get $0.10 of free credit.
Pair Qwen 4 Max with a media API
Use Qwen4-Max for the reasoning, then generate the images, video, music and 3D models it plans on Synexa. Pay per request, start with $0.10 free credit.