# Qwen4-Max Preview review: Coding tests, GameFeel scores, pricing and setup [Read this article on Offloop](https://offloop.org/learn/qwen-4-max-preview-test) Yifei Qwen has opened access to Qwen4-Max Preview, the first publicly available flagship multimodal model in the Qwen4 family. It supports code generation, complex reasoning and agent tasks. The first day of testing produced a mixed picture. One developer generated a working interactive 3D lighthouse island with a single prompt. In the independent GameFeel RE test, Qwen4-Max Preview scored 10.3, below the 12.1 scored by its predecessor, Qwen3.8-Max. Public benchmarks are still too limited to judge its overall coding ability. This review covers the published specs, GameFeel scores, three coding examples and how to try the model in Qwen Chat, Qoder and through the API. ## What is Qwen4-Max Preview? Qwen4-Max Preview is an early release of Qwen’s flagship Qwen4 model for coding, office work and multistep agent tasks. It accepts text, images and video and supports web search, tool calling and structured output. Its 1-million-token context window can accommodate large amounts of code, documents and screenshots in the same task. [Official Qwen4-Max Preview documentation](https://docs.qwencloud.com/developer-guides/getting-started/latest-model) ### Qwen4-Max Preview specifications Model specifications | Item | Specification | | ------------------------ | ------------------------------------------------------- | | Model name | Qwen4-Max Preview | | API model ID | qwen4-max-preview | | Model alias | qwen4.0-max-preview | | Context window | 1 million tokens | | Maximum output | 393,216 tokens shared by reasoning and the final answer | | Reasoning effort | low, medium, xhigh | | Default reasoning effort | xhigh | | API protocols | OpenAI Chat Completions, Responses, Anthropic Messages | | Preview dates | November 11–25, 2026 | The API preview runs from 10:00 on November 11 until the end of November 25, Beijing time. This preview model will then be retired. Qoder offered an anonymous test model called Qoduck for four days before replacing it with Qwen4-Max Preview in the model picker. Check the dates on earlier demos, which may have used a Qoduck build. [Qoder launch announcement](https://x.com/qoder_ai_ide/status/2120343731673108936) ![Qoder’s announcement of Qoduck and Qwen4-Max Preview](https://offloop.org/editorial/learn/qwen-4-max-preview-test/qoder-qoduck-original.jpg) Qoder’s original X announcement advertises 80% off. The post separately lists 0.1× Credits. ## How does Qwen4-Max Preview score on benchmarks? As of November 11, 2026, Qwen4-Max Preview had no entry on the public Artificial Analysis, Arena Text or Arena WebDev leaderboards. The official launch materials did not provide a full set of benchmark results. One independent result comes from GameFeel Benchmark. ### GameFeel RE: 10.3, below Qwen3.8-Max ![GameFeel RE scores: Qwen4-Max Preview 10.3, Qwen3.8-Max 12.1, Qwen3.8-Flash-Next 13.0, Qwen4-27B Early 14.9 and Qwen4-Flash Early 19.9](https://offloop.org/editorial/learn/qwen-4-max-preview-test/gamefeel-re.png) Independent GameFeel RE scores, redrawn from the published results. The scale is out of 100. This view shows 0–20. GameFeel RE scores | Model | GameFeel RE score | | ------------------ | ----------------- | | Qwen4-Max Preview | 10.3 | | Qwen3.8-Max | 12.1 | | Qwen3.8-Flash-Next | 13.0 | | Qwen4-27B Early | 14.9 | | Qwen4-Flash Early | 19.9 | GameFeel measures abilities related to game experience. RE (Reverse Evaluation) is one of its seven categories and tests how well models assess proposed game changes, especially their effect on balance. Qwen4-Max Preview trailed Qwen3.8-Max and several early Qwen4 models on RE. That one result cannot establish how the preview compares with its predecessor at code generation overall. [Original GameFeel results post](https://x.com/follistatindev/status/2120350029915230424) · [Score breakdown](https://x.com/bebuu_ai/status/2120404110348587294) · [Benchmark methodology](https://github.com/kodasdream/gamefeel-benchmark#how-the-benchmark-works) ![Original GameFeel RE chart showing Qwen4-Max Preview at 10.3](https://offloop.org/editorial/learn/qwen-4-max-preview-test/gamefeel-re-original.jpg) Original GameFeel RE chart posted by @follistatindev on X. The code people shared on launch day offers a closer look at what the model can build and where it struggles. ## How well does Qwen4-Max Preview code? Three hands-on examples After the preview opened, developers posted websites, small games and SVG animations on X. Some outputs ran on the first attempt. Animation details and gameplay still varied across the examples. ### 1. An interactive 3D lighthouse island from one prompt Developer AJ tried the same 3D lighthouse island prompt with several models. Qwen4-Max Preview generated an interactive scene in one attempt and added working controls that the original prompt had not requested. AJ also reported that Qwen4-Max Preview used fewer tokens than DeepSeek V4.1 Flash on his version of the task. The result shows what the model can do with a standalone interactive page. Token use on one task offers a limited basis for comparing coding costs. [Original 3D lighthouse island test](https://x.com/ItsmeAjayKV/status/2120415237950325148) ![The 3D lighthouse island and its control panel on the right](https://offloop.org/editorial/learn/qwen-4-max-preview-test/lighthouse-island-frame.jpg) The lighthouse and its controls at 00:02 in @ItsmeAjayKV’s original video. ### 2. A pelican on a bicycle, with improved SVG animation and rough details Developer Leo tried the pelican-on-a-bicycle SVG animation again in Qoder. He saw some improvement over the earlier Qoduck build, although he found the difference small. Another tester pointed out that the pelican’s feet did not line up with the bicycle pedals. An AI-generated animation can look complete while still getting the coordination between moving parts wrong. [Leo’s SVG animation test](https://x.com/leo114119/status/2120345166997324170) · [Original post showing the pedal detail](https://x.com/ai_funss/status/2120348790997844109) ![The cycling pelican’s feet and the positions of the bicycle pedals](https://offloop.org/editorial/learn/qwen-4-max-preview-test/pelican-pedals-frame.jpg) A close look at the feet and pedals at 00:04 in @ai\_funss’s original video. ### 3. Crossy Road, with results close to Claude Opus 5.5 Developer Mark Santos gave Qwen4-Max Preview and Claude Opus 5.5 Medium the same prompt. `build the crossy road game` Both models produced an HTML game. Mark downloaded and compared them and gave Claude Opus 5.5 a slight edge. [Original Crossy Road comparison](https://x.com/markksantos/status/2120446429718086142) ![Side-by-side comparison of Crossy Road games generated by Qwen4-Max Preview and Claude Opus 5.5 Medium](https://offloop.org/editorial/learn/qwen-4-max-preview-test/crossy-road-frame.jpg) Qwen4-Max Preview on the left and Claude Opus 5.5 Medium on the right, at 00:20 in @markksantos’s original video. These examples show that Qwen4-Max Preview can build standalone websites, 3D scenes and small games. The quality varied by task. Most public examples involve building a standalone frontend in one attempt. Large repository changes, difficult debugging and long-running agent workflows need more testing. ## How do you use Qwen4-Max Preview? There are three main ways to try it: Qwen Chat, Qoder and API access through QwenCloud Token Plan. ### 1. Try it for free in Qwen Chat The quickest way to try it is [Qwen Chat](https://chat.qwen.ai/). Sign in and choose Qwen4-Max Preview in the model picker. To test a website or a small game, switch to Agent mode and select the model there. A free account was reported to have used the model on launch day. Check your account for the available allowance. [Free account test](https://www.reddit.com/r/Qwen_AI/comments/1xtayah/tried_qwen4_preview_in_completely_new_account_in/) · [Qwen launch announcement](https://x.com/Alibaba_Qwen/status/2120335993643684315) ![Qwen’s original Qwen4 teaser image](https://offloop.org/editorial/learn/qwen-4-max-preview-test/qwen-announcement-original.jpg) Original teaser posted by Qwen on X. To test its coding ability yourself, try this prompt. > Build a single-file HTML expense dashboard using sample data. Let me filter by category, switch months and reset the filters. Make it work at mobile widths too. Open the result and try the filters, month switcher and reset button. Then narrow the browser window to check the layout. ### 2. Use Qoder in your development environment Use Qoder to let the model read and edit files in a project. Start a new task, click the current model name at the bottom of the composer and choose `Qwen4-Max-Preview` from the list. Qoder’s launch announcement said the preview was open across its plans. The Credits rate can vary, so check the rate in the model picker before you send a task. [Qoder model selector guide](https://docs.qoder.com/qoder/model-selector) ### 3. Set up API access through QwenCloud Token Plan Use QwenCloud Token Plan to access Qwen4-Max Preview in developer tools that support the OpenAI or Anthropic protocols. Here is how to set it up in Qwen Code Desktop. 1. Subscribe to Token Plan in the QwenCloud console and get a plan-specific API key. 2. Go to More → Settings → Models → Add Model. 3. Select QwenCloud → Token Plan → Singapore, enter the API key and choose `qwen4-max-preview`. QwenCloud’s international service provides the following compatible endpoints. QwenCloud international base URLs | API protocol | Base URL | | ------------ | ------------------------------------------------------------ | | OpenAI | https\://token-plan.maas.qwencloudapi.com/compatible-mode/v1 | | Anthropic | https\://token-plan.maas.qwencloudapi.com/apps/anthropic | Token Plan Individual is for interactive use in coding and agent tools. Its terms exclude application backends, batch processing and automation scripts. [Qwen Code setup guide](https://docs.qwencloud.com/developer-guides/clients-and-developer-tools/qwen-code) · [Token Plan quickstart](https://docs.qwencloud.com/token-plan/personal/token-plan-personal-quickstart) · [Individual plan terms](https://docs.qwencloud.com/token-plan/personal/token-plan-personal-overview) ## How much does Qwen4-Max Preview cost? Is it free? You can try Qwen4-Max Preview for free in Qwen Chat. API access requires a QwenCloud Token Plan subscription. Pay-as-you-go token billing is unavailable during the preview. Here are the individual plan prices on QwenCloud’s international service. QwenCloud international individual plan pricing | Plan | Monthly price | Monthly Credits | | --------- | -------------------------------------- | --------------- | | Lite | $6, limited-time $3 for new customers | 20,000 | | Essential | $10, limited-time $5 for new customers | 45,000 | | Standard | $18 | 80,000 | | Pro | $43 | 200,000 | [Official QwenCloud Token Plan pricing](https://docs.qwencloud.com/token-plan/personal/token-plan-personal-overview) There are promotional Credits rates during the preview, and the published rates vary by platform. QwenCloud’s international documentation lists a promotional rate as low as 20% of standard Credits, or as low as 8% from 22:00 to 08:00 Beijing time. Alibaba Cloud also announced an overnight offer. Qoder’s launch post mentions a 0.1× Credits rate. Its model documentation lists 0.5×. Check the Credits rate where you plan to use the model. Different platforms deduct Credits at different rates. [Alibaba Cloud offer announcement](https://x.com/alibaba_cloud/status/2120383278297264543) · [Qoder launch offer](https://x.com/qoder_ai_ide/status/2120343731673108936) ![Alibaba Cloud’s Qwen4-Max Preview Token Plan offer image](https://offloop.org/editorial/learn/qwen-4-max-preview-test/token-plan-offer-original.jpg) Alibaba Cloud’s original X announcement lists 80% off for a limited time and 88% off at night. ## Is Qwen4-Max Preview open source? As of November 11, 2026, **Qwen4-Max Preview’s model weights were unavailable**, so you could not download it for local deployment. Qwen also had not announced official release dates or weight downloads for Qwen4-27B, Qwen4-Flash or Qwen4-Plus. [Qwen model release notes](https://docs.qwencloud.com/changelog/models) The claim that Qwen4 outscored Claude Opus 5.5 on gaming benchmarks refers to a specific early model. Qwen4-Flash Early averaged 49.9 across GameFeel’s seven categories, compared with 47.2 for Claude Opus 5.5 XHigh. Those scores came from the early Flash build. [Original GameFeel test post](https://x.com/follistatindev/status/2119220233751671186) ![Original GameFeel overall chart showing Qwen4-Flash Early at 49.9 and Claude Opus 5.5 XHigh at 47.2](https://offloop.org/editorial/learn/qwen-4-max-preview-test/gamefeel-overall-original.jpg) Original GameFeel overall chart posted by @follistatindev on X. For a related model with published weights, see [Qwen3.8-Flash-Next](https://huggingface.co/Qwen/Qwen3.8-Flash-Next). The Qwen team describes it as an experimental preview of the architecture behind Qwen4. ## Is Qwen4-Max Preview worth trying? It is worth trying for websites, interactive pages and small games. The launch-day examples include an interactive 3D scene from a single prompt, and you can try it in Qwen Chat. For production development, those examples leave open how the model compares with the previous Qwen generation or competitors such as Claude. The 10.3 in GameFeel RE shows a weak result on that task. The public frontend demos show what it can generate. Broader benchmarks and project tests would help assess both capabilities. The API preview ends on **November 25, 2026**. Teams using it for daily development should plan how to switch models once the preview is retired. ## Frequently asked questions about Qwen4-Max Preview ### Can I try Qwen4-Max Preview for free? Look for Qwen4-Max Preview in the Qwen Chat model picker. A free account used it on launch day. Check the available allowance in your account. ### Does the Qwen4-Max Preview API offer pay-per-token billing? During the preview, QwenCloud offers API access through a Token Plan subscription. Calls consume plan Credits. Pay-per-token billing is unavailable. ### How large is the Qwen4-Max Preview context window? The official context window is 1 million tokens. Reasoning and the final answer share a maximum output budget of 393,216 tokens. ### Has Artificial Analysis or Arena scored Qwen4-Max Preview? As of November 11, 2026, Artificial Analysis, Arena Text and Arena WebDev had no public score for this model. The independent GameFeel RE score of 10.3 covers the game reverse-evaluation category. ### Can I download Qwen4-Max Preview and run it locally? As of November 11, 2026, Qwen had not released weights for the preview or published weight downloads for Qwen4-27B, Flash or Plus. ### When does the Qwen4-Max Preview API preview end? The API preview runs through the end of November 25, 2026 (UTC+8). QwenCloud will retire this preview version when the period ends. ## Sources * [\[1\] Official Qwen4-Max Preview documentation](https://docs.qwencloud.com/developer-guides/getting-started/latest-model) * [\[2\] Qoder launch announcement](https://x.com/qoder_ai_ide/status/2120343731673108936) * [\[3\] Original GameFeel results post](https://x.com/follistatindev/status/2120350029915230424) * [\[4\] Score breakdown](https://x.com/bebuu_ai/status/2120404110348587294) * [\[5\] Benchmark methodology](https://github.com/kodasdream/gamefeel-benchmark#how-the-benchmark-works) - [\[6\] Original 3D lighthouse island test](https://x.com/ItsmeAjayKV/status/2120415237950325148) - [\[7\] Leo’s SVG animation test](https://x.com/leo114119/status/2120345166997324170) - [\[8\] Original post showing the pedal detail](https://x.com/ai_funss/status/2120348790997844109) - [\[9\] Original Crossy Road comparison](https://x.com/markksantos/status/2120446429718086142) - [\[10\] Qwen Chat](https://chat.qwen.ai/) - [\[11\] Free account test](https://www.reddit.com/r/Qwen_AI/comments/1xtayah/tried_qwen4_preview_in_completely_new_account_in/) - [\[12\] Qwen launch announcement](https://x.com/Alibaba_Qwen/status/2120335993643684315) - [\[13\] Qoder model selector guide](https://docs.qoder.com/qoder/model-selector) - [\[14\] Qwen Code setup guide](https://docs.qwencloud.com/developer-guides/clients-and-developer-tools/qwen-code) - [\[15\] Token Plan quickstart](https://docs.qwencloud.com/token-plan/personal/token-plan-personal-quickstart) - [\[16\] Individual plan terms](https://docs.qwencloud.com/token-plan/personal/token-plan-personal-overview) - [\[17\] Alibaba Cloud offer announcement](https://x.com/alibaba_cloud/status/2120383278297264543) - [\[18\] Qwen model release notes](https://docs.qwencloud.com/changelog/models) - [\[19\] Original GameFeel test post](https://x.com/follistatindev/status/2119220233751671186) - [\[20\] Qwen3.8-Flash-Next](https://huggingface.co/Qwen/Qwen3.8-Flash-Next) November 11, 2026