Can Qwen4 Max Close The Gap? A Nine-Point Gain At Lower Cost
AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Artificial Analysis gave Alibaba’s Qwen4 Max an Intelligence Index score of 43, nine points above the previous Qwen flagship, and estimated its index task cost at $3.27 versus $10.02. The launch-day results place it above the comparable-model median, though it ranked ninth among 10 models on the Qwen4 Max page; speed data was not available.

Alibaba’s Qwen4 Max scored 43 on Artificial Analysis’s Intelligence Index, nine points above the previous Qwen flagship, while its estimated cost per index task was lower. The independent benchmark, published on November 16, places the new model above the comparable-model median but near the bottom of the models listed on its page; launch-day speed data was unavailable.

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

Artificial Analysis lists Qwen4 Max at $3.27 per index task, compared with $10.02 for Qwen3.8 Max (0902). The company’s listed API prices are $1.50 per million input tokens and $4.50 per million output tokens. These figures describe the benchmark’s estimated task costs and published token prices; they do not establish what every customer will pay for a particular workload.

On the same index, Qwen4 Max’s score of 43 is above the reported median of 41 for models at a similar price. Yet it ranks ninth among the 10 models shown on the Qwen4 Max page, whose top score is 65. Those comparisons describe its placement on that page and index, not performance on every task.

Alibaba introduced Qwen4 Max as the first model in its Qwen4 family. The company says it has a 2.4-trillion-parameter backbone, 71 billion parameters activated per token and a one-million-token context window. It is available through Alibaba’s QwenCloud API, Token Plan subscription and apps including QwenWork, Qoder and QwenStudio. The model’s weights are not publicly downloadable.

At a glance
reportWhen: Published November 16, 2026; launch-day…
The developmentArtificial Analysis published its first Qwen4 Max benchmark results on November 16, reporting a nine-point gain over its predecessor and a lower estimated cost per index task.
Bar chart titled Ninth of ten on its own page, nine points above its predecessor. Artificial Analysis Intelligence Index for the 10 models on the Qwen4 Max page: Claude Fable 5.2 (max) 65, Gemini 4 Carbon (high) 60, Claude Opus 5.5 (max) 59, GPT-6 Astra (max) 57, GPT-6.1 Sol (max) 55, Muse Wonder Full (max) 50, Kimi K3.1 Pro (open weights) 46, Muse Wonder (max) 44, Qwen4 Max 43 (highlighted), Grok 4.7 (xhigh) 35. Below a dashed line: the predecessor Qwen3.8 Max (0902) 34
Qwen4 Max is ninth of the ten models on its Artificial Analysis page and nine points above its predecessor. Source: Artificial Analysis, model pages for Qwen4 Max and Qwen3.8 Max (0902), retrieved November 16, 2026. Chart: Thorsten Meyer AI.

Lower Benchmark Costs, Ongoing Trade-offs

The results give buyers an initial outside comparison of the new flagship against its predecessor. The combination of a higher index score and a lower estimated task cost may make Qwen4 Max worth evaluating for organizations that use the benchmark’s kinds of tasks. It does not, by itself, show that the model will be cheaper or more capable for a given company’s workload.

Its ranking also puts the nine-point gain in perspective. Qwen4 Max improved over Qwen3.8 Max, but the page’s reported top score of 65 is well above its 43. The model’s position will matter to teams choosing among frontier systems, while cost, task quality, latency and access requirements may matter more for specific deployments.

Access is another practical factor. Because the weights are closed, customers depend on Alibaba’s services to use the model. The company says smaller Qwen4 models will be open-sourced, but has not given a release date in the cited material.

Products Worth Considering

Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite

Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

As an affiliate, we earn on qualifying purchases.

Alibaba’s Claims and Outside Results

Alibaba’s launch blog reports a score of 52.69 on its Work360 benchmark, up from 47.74 for Qwen3.8 Max. The company also says average cost per task on Work360 fell from $4.43 to $1.67, a 62% reduction. These are vendor-reported results from Alibaba’s own evaluation, so they are distinct from Artificial Analysis’s independent index and cost estimate.

The launch post describes Qwen4 Max as a multimodal model that can process text, images and video, and says reasoning is enabled by default at high effort. Alibaba also reports gains on coding and secure-coding tests, along with examples from manufacturing and other professional settings. Those figures and demonstrations come from the company; the cited material does not establish that they were independently reproduced or obtained under controlled conditions.

Artificial Analysis published its first Qwen4 Max measurements on the launch date. Its model pages for Qwen4 Max and Qwen3.8 Max (0902), retrieved November 16, are the cited basis for the independent scores and task-cost comparison. A limited preview reportedly opened on November 11, according to The Next Web.

“Your new teammate has arrived.”

— Alibaba’s Qwen team

Speed, Workloads and Open Weights

Speed and latency had not been measured by Artificial Analysis when it published its launch-day results. The available index and cost figures therefore do not show how quickly Qwen4 Max responds or how it performs under different traffic conditions.

It is also not yet clear how well the model’s benchmark gains translate to individual customer tasks. Alibaba’s Work360 results and case studies are vendor-reported, while the independent score summarizes performance through a particular index. The supplied material does not provide enough detail to resolve how results vary by workload or deployment.

Finally, Alibaba has not published Qwen4 Max’s weights. The company says smaller models in the family will be open-sourced, but the timing and terms are unspecified in the cited launch information.

More Measurements and Family Releases

Further Artificial Analysis results, especially speed and latency measurements, could help clarify how Qwen4 Max compares in practical use. Independent evaluations on specific workloads would also help test whether the benchmark and cost advantages hold for particular applications.

Alibaba has said smaller Qwen4 models will be open-sourced, but has not announced when they will be available. For now, Qwen4 Max can be accessed through Alibaba’s listed API, subscription and apps, with performance and total cost depending on how customers use those services.

Key Questions

What was Qwen4 Max’s independent benchmark score?

Artificial Analysis scored it 43 on its Intelligence Index on November 16, 2026.

How much did the score improve?

The score was nine points higher than the 34 recorded for Qwen3.8 Max (0902) on the same index.

How does its estimated task cost compare?

Artificial Analysis estimated $3.27 per index task for Qwen4 Max and $10.02 for Qwen3.8 Max (0902). Actual costs can vary by workload.

Is Qwen4 Max open source?

No. Its weights are not publicly downloadable. Alibaba says smaller Qwen4 models will be open-sourced, without a date in the cited material.

Are speed results available?

Artificial Analysis had no speed or latency data for Qwen4 Max on launch day.

Source: Thorsten Meyer AI

You May Also Like

One upload in. A whole channel’s worth of content out.

ChannelHelm v1.5 adds A/B testing, thumbnail learning and retention-based clip selection for creators using one video across platforms.

Inside MentalHealthBench: A New Perspective On AI And Mental Health

OpenAI says MentalHealthBench will evaluate AI responses to mental health conversations. Its design and real-world relevance await independent review.

CTOs Are Escaping

Senior CTOs are leaving traditional roles for hands-on positions at Anthropic, signaling a shift in tech power towards AI model development and experimentation.

Which AI Boss Would You Trust With the Company? Take the Test

Can you spot an AI manager by its decisions? Firmulate turns 242 audited choices into a quiz about judgment, discipline and follow-through.