Close Menu
  • AI
  • Content Creation
  • Tech
  • Robotics
AI-trends.todayAI-trends.today
  • AI
  • Content Creation
  • Tech
  • Robotics
Trending
  • Lean and Mathlib Communities Certify Claude-Assisted Riemann Proof – Unite.AI
  • AWS Routes HyperPod Inference Throughout Areas From One Endpoint – Unite.AI
  • DailyAI Briefing: A Lean-checked Riemann proof, plus 4 extra
  • Anthropic Shares Drop 8.5% in Second Session Throughout AI Selloff – Unite.AI
  • OpenAI Ships GPT-Stay-1.1 for Voice Brokers, Cuts Idle Value to $0.02/Min
  • Copywriting: The Full Information | Copyblogger
  • AI Governance in 2027: Accountability Strikes Into the Workflow – Unite.AI
  • Reply Engine Optimization (AEO): The Full Information
AI-trends.todayAI-trends.today
Home»Tech»Qwen introduces Qwen3.7: A reasoning agent model with a 1M token context window

Qwen introduces Qwen3.7: A reasoning agent model with a 1M token context window

Tech By Gavin Wallace22/05/202610 Mins Read
Facebook Twitter LinkedIn Email
Samsung Researchers Introduced ANSE (Active Noise Selection for Generation): A
Samsung Researchers Introduced ANSE (Active Noise Selection for Generation): A
Share
Facebook Twitter LinkedIn Email

The majority of AI models aren’t designed to execute in multiple steps and for a sustained period. For tasks such as running hundreds of code iterations or chaining calls to tools over several hours, a model with a training focus that is different from the current AI models will be required.

Qwen3.7 Max was officially announced by the Qwen team at 2026 Alibaba Cloud Summit, on 20 May. Two preview versions of Qwen3.7 appeared quietly on Arena AI leaderboard without any press release or official API announcement.

The release of two preview models at the same time

Alibaba previewed two models simultaneously: Qwen3.7-Max-Preview and Qwen3.7-Plus-Preview. According to LM Arena, they ranked respectively 13th and 16th worldwide in terms of text capability.

Alibaba is the 6th lab in Text Arena. Qwen3.7 Max-Preview was ranked 13 overall. Qwen3.7 Plus-Preview placed #16 in Vision Arena. Alibaba is the #5 laboratory. Model rank and lab rank are two separate numbers.

The Qwen3.7 Plus-Preview preview is a balanced high-performance version that focuses on logic and reasoning. Its toolchain will be opened gradually in the future. This version supports multimodal input and visuals. Qwen3.7 is the flagship text-only reasoning. The Qwen3.7 Max model is covered in this article, since it’s the one Alibaba officially announced.

Why is Qwen3.7 Max Designed for?

Alibaba Qwen’s team describes Qwen3.7 Max as their most sophisticated and comprehensive agent to date. Model is proprietary, closed-weight. It can handle coding and bugging, automated office workflows, and large-scale tasks with hundreds, if not thousands, of steps.

Long-Term Thinking Mode

The Qwen3.7 Max is a logic model. The model generates a chain of thought first — an internal sequence of steps where it plans, checks its work, and corrects course before committing to a final answer. On interfaces like Qwen Chat, this shows up as a ‘Thinking’ mode you can switch on to see the model’s reasoning trace.

The output tokens of reasoning models are significantly higher than those produced by standard completions. Artificial Analysis evaluated the Intelligence Index using Qwen3.7. Max. It generated around 97 million tokens, compared to an average of 24 million For models based on this benchmark. This overhead increases latency for short tasks or simpler ones without increasing output quality. The extended-thinking model is most useful for multi-step plans, code refactoring or large agent chains.

Contextual Window

Model features 1M context windows, up to 256K in Qwen3.6. The model supports only text input and outputting. The pricing has yet to be announced. Alibaba Cloud charges $1.30/$7.80 for each million tokens of input/output.

One context window that can store a thousand tokens is enough to hold an entire mid-sized repository of code or many documents. As the context window grows, models tend to become less reliable. Qwen3.7 Max is still not available for independent long context testing.

Benchmark Results

The Artificial Analysis Intelligence Index ranked Qwen3.7 Max at 56.6, ranking it as the fifth best overall. It is up 4.8 points from its predecessor Qwen3.6 Max Preview (51,8) and ahead of Google Gemini 3.5 Flash’s (55.3). GPT-5 (55.2), Claude Opus 4.7 (57.3), and Gemini 3.1 Preview (57.2), still top the list.

The Intelligence Index v4.0 contains ten assessments, such as GDPval-AA (Average), Terminal-Bench-Hard, SciCode, AA Omniscience, Humanity’s Last Exam and GPQA.

https://qwen.ai/blog?id=qwen3.7

The improvements over Qwen3.6 are not consistent. Index improvements are concentrated primarily in three areas: scientific reasoning and agentic capabilities. CritPt jumped from 3.7% to 134%, Humanity’s last exam jumped by 9.2 percentage points and Terminal Bench Hard rose 6.9 percentage points. GDPval-AA gained 42 Elo (from 1504 to 1546). Other benchmark scores are mostly flat in comparison to Qwen3.6 Maximum Preview.

The Index is a complex index. In AA Omniscience, the raw accuracy for Qwen3.7 Max actually fell 7.6 percentages points (from 36.7% to 30%.1) while its hallucination ratio dropped 21.3 percentage points (from 42.2% to 22.9%). Model is choosing what to say “I don’t know” More facts are not necessarily better. Its rate of attempts fell from 67.3 to 48.0%. That was the lowest attempt rate among all frontier models in this comparison. The AA Omniscience Benchmark rewards correct answers but penalizes hallucinations. There is no penalty if you refuse to answer. If your use case relies on broad facts being remembered, then this limitation is important to consider when testing against workload.

Qwen3.7 Max-Preview was ranked 13th overall in Text Arena with an Elo of 1475. The top ten categories are: #7 for Math, #9 for Expert Prompts and IT and #10 for Coding.

All benchmarks are tentative. The model carries a ‘Preview’ mode, indicating Alibaba considers it an early build.

Agentic Performance — Internal Test

Alibaba’s internal test of a chip platform saw the model perform more than 1,000 calls to tools and make iterative changes in code. This was done for optimizing a kernel. Alibaba said the inference process increased speed roughly by 10x when compared to the old version.

Marktechpost’s Visual Explainer






The first slide of six

What exactly is Qwen3.7 Max?

Alibaba has developed a proprietary model of reasoning for multi-step automated tasks and long-horizon agents.

Contextual Window

1 million tokens — enough to fit a full mid-sized code repository in a single request.

Justification Model

Before deciding on a conclusion, use the extended-thinking (chain-of thought) mode.

Input/Output

Text out, no image input. This model does not support image input.

The API string

Use qwen3.7-max When calling through Alibaba Cloud Model Studio.

Apache-compatible API
OpenAI & Anthropic spec
Preview — no open weights yet

This is Slide 6 of 6.

Chat Interface Quick Start

This is the fastest and easiest way to try Qwen3.7.Max without an API key.

  • 1

    Qwen Chat

    Navigation to the site chat.qwen.ai You can create an account for free.

  • 2

    Please select the appropriate model

    Choose the desired model in the Model Selector Dropdown. Qwen3.7-Max. This may be a. Qwen3.7-Max-Preview During the pre-view period,

  • 3

    Think Mode

    Switch on The Thinking Mode This will activate the chain-of-thought reasoning and show you how to use it in chat. It activates the chain of thought reasoning, and displays the model’s reasoning trail before the final response.

  • 4

    You can send your message at any time

    Type your query. If you are working on a complex task, it is best to be as specific as possible about the steps, restrictions, and output format.

💡

Test with the hardest prompts you can find. Simple prompts are not as effective at revealing model quality. Multi-step problems in math, complex requests for refactoring, and unclear expert questions can reveal more than simple ones.

The third slide of six

Access to API

Qwen3.7 Max is compatible with OpenAI API Specifications as well as Anthropic API Specifications. It can be easily integrated into pipelines.

OpenAI-compatible Python call

OpenAI import OpenAI

OpenAI client = OpenAI
    apI_key="YOUR_DASHSCOPE_API_KEY",
    base_url="https://dashscope-intl.aliyuncs.com/compatible-mode/v1"
)

response = client.chat.completions.create(
    model="qwen3.7-max",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user",   "content": "Explain chain-of-thought reasoning."}
    ]
)

print(response.choices[0].message.content)

ℹ️

You can get your API Key from Alibaba Cloud Model Studio (DashScope). This is the base URL of international access. dashscope-intl.aliyuncs.com.

⚠️

The pricing for Qwen3.7 Max has yet to be announced. Qwen3.6 max preview was priced at $7.80 / $1.30 per million tokens.

This is Slide 4, of 6.

Understand Thinking Mode

The Thinking Mode layer is the chain of thought reasoning for the model. This determines the way in which the model will approach a particular problem and generate a solution.

What to do with it

Complex math proofs and long agents task chains. Uncertain problems requiring step by step planning.

Why you should skip the ad?

Quick lookups and quick rewrites.


Use extra_body to think via API

response = client.chat.completions.create(
    model="qwen3.7-max",
    messages=[{"role":"user","content":"Your prompt here"}],
    extra_body={"enable_thinking": True}
)

💡

Qwen3.7 generated about 97M tokens based on Artificial Analysis benchmarks compared to the average 24M of comparable models. Each thinking token adds to latency and cost — use thinking mode selectively.

This is Slide 5.

Tasks that are both long-term and agentic

Qwen3.7 Max is built to perform long-running, autonomous tasks loops. Alibaba tested it internally and found that the system was capable of running up to 350 hours autonomously, executing 1,000+ calls for tools.

  • 1

    Define your tools in a clear way

    OpenAI Standard: Tool definitions are now available. The following are some examples of how to use parameter. It is possible to invoke functions and tools iteratively using the model.

  • 2

    You can use the 1M window to display context.

    Contextualize the full history of tasks, previous tool outputs and current code. Trim aggressively when the full context is not needed — every token is billed.

  • 3

    Targeting the last answer

    A standard completion is shorter and less variable. If you are writing tests for students, make sure to focus more on the end result than the exact words of the thought trace.

  • 4

    Useful Cases

    The optimisation of the kernel, loops for debugging code, automation of office workflows and pipelines that include multiple steps with iterative validation.

⚠️

Alibaba only tested the figures for 35 hours and 1000+ tools. These claims are not independently verified.

This is Slide 6 from 6.

Know the Limitations

By understanding these limitations, you can save time on debugging and set realistic expectations.

No image input

Qwen3.7Max only supports text input. Qwen3.7 Plus-Preview is the best choice for multimodal tasks.

AA-Omniscience Abstention

On the AA Omniscience standard, the model’s try rate fell from 67.3% (on average) to 48.0%. It abstains more and hallucinates less — but its raw factual recall also dropped. Check your knowledge recall.

Preview status

The model currently carries a — Preview suffix. Before a stable release, benchmark scores, behavior, and pricing may change. There will be no open weight version available after May 2026.

Long-term reliability

It is important to note that a 1M token window in the context of Qwen3.7-Max represents a limit, and not a promise. Qwen3.7 Max does not have independent long context testing. Validate retrieval performance on your workload.

ℹ️

For the latest model updates, check the offIcial Qwen blog at qwen.ai/blog Alibaba Cloud Model Studio documentation.

The Key Takeaways

  • Alibaba has released two Qwen3.7 models as previews: Plus (multimodal) and Max (text/reasoning).
  • Qwen3.7-Max scored 56.6 on the Artificial Analysis Intelligence Index, ranking #5 overall — a 4.8-point gain over Qwen3.6 Max Preview.
  • Text only; no images.
  • On AA-Omniscience, raw accuracy dropped while abstention rose — worth testing for knowledge-recall use cases.
  • Alibaba has only conducted internal tests on the model, and no external verification.

Check out the Technical details. You can also find out more about the following: Docs.  Also, feel free to follow us on Twitter Join our Facebook group! 150k+ ML SubReddit Subscribe now our Newsletter. Wait! Are you using Telegram? now you can join us on telegram as well.

You can partner with us to promote your GitHub Repository OR Hugging Page OR New Product Launch OR Webinar, etc.? Connect with us


ces x
Share. Facebook Twitter LinkedIn Email
Avatar
Gavin Wallace

Related Posts

OpenAI Ships GPT-Stay-1.1 for Voice Brokers, Cuts Idle Value to $0.02/Min

21/11/2026

Meet Pondering Machines Inkling 2 Mini: 26B Looped MoE Runs on a 32GB Mac

21/11/2026

Baseten Releases Bison-1 and Groundhog-1, Topping Choice Mannequin Index

21/11/2026

RadixArk Ships DaisyRose-27B, a 27B Agentic Coding Mannequin Skilled With Reinforcement Studying Solely

20/11/2026
Top News

The AI tool will tell you to stop slacking off

‘This Is Not a Drill’: Essential PHP Vulnerabilities Are Setting the Web on Fireplace

A Flaw in OpenAI’s Codex Put 97M Month-to-month Downloads at Threat

Then, is it? “AI bubble” About to burst late in 2025 or even 2026?

What if I find an AI agent that is worth the risk?

Load More
AI-Trends.Today

Your daily source of AI news and trends. Stay up to date with everything AI and automation!

X (Twitter) Instagram
Top Insights

Reid Hoffman believes doctors should consult AI to get a second opinion

30/04/2026

Amazon Is Making an AI-Animated ‘Good Advice Cupcake’ TV Show. The original creator is furious

29/05/2026
Latest News

Lean and Mathlib Communities Certify Claude-Assisted Riemann Proof – Unite.AI

21/11/2026

AWS Routes HyperPod Inference Throughout Areas From One Endpoint – Unite.AI

21/11/2026
X (Twitter) Instagram
  • Privacy Policy
  • Contact Us
  • Terms and Conditions
© 2026 AI-Trends.Today

Type above and press Enter to search. Press Esc to cancel.