Qwen3.6 Max Preview

qwen3.6-max-preview

Get API keyAPI reference

Qwen3.6 Max Preview is built on a hybrid architecture that combines efficient linear attention with sparse mixture-of-experts routing, enabling strong scalability and high-performance inference. Compared to the 3.5 series, it brings significant advancements in agentic coding, front-end development, and overall reasoning, with a markedly improved "vibe coding" experience. The model excels at complex tasks such as 3D scenes, games, and repository-level problem solving, scoring 78.8 on SWE-bench Verified. It represents a major leap in both text-only and multimodal capabilities, reaching the level of leading state-of-the-art models.

Context window1.0M

Provider通义千问

Released2026/04/22

Playground

Pricing

The same model is available through multiple service channels — choose based on latency, reliability and cost.

Prices in $ / 1M tokens

To pick a channel, add a provider field to the request body, for example "provider": { "channel": "direct" }. Valid values are direct / stable / economical; omit it to use the default channel.

Direct

Direct upstream connection — best when you need native behavior and the full context window.

Input context	Input	Output
≤ 128K	1.32/M	7.94/M
> 128K	2.21/M	13.24/M

Overview

Input

Text Image

Output

Text

Qwen3.6 Max Preview API: A Hybrid Architecture Flagship for Frontend and Repository-Level Problems

Qwen3.6 Max Preview is built on a hybrid architecture that combines efficient linear attention with sparse mixture-of-experts routing, delivering strong scalability and high-performance inference. Compared to the 3.5 series, it brings significant improvements in agentic coding, frontend development, and overall reasoning, with a notably improved "vibe coding" experience.

This model excels at complex tasks such as 3D scenes, games, and repository-level problem solving, with standout results on real-world engineering benchmarks like SWE-bench. If your work centers on frontend and interactive development, Qwen3.6 Max is well worth trying.

SeaWhale AI provides Qwen3.6 Max Preview through an OpenAI-compatible API, supporting multimodal input, tool calling, and streaming output.

Get API Key · Model ID: qwen3.6-max-preview

---

Why Choose Qwen3.6 Max Preview

• Hybrid architecture — Linear attention + sparse MoE routing, balancing scalability and inference efficiency
• Frontend development strength — Major improvements in frontend and "vibe coding" experience over the 3.5 series
• 3D scenes & games — Outstanding capability in generating complex interactive content
• Repository-level problem solving — Excellent performance on real-world engineering tasks like SWE-bench
• 1M token context — Handle large projects in a single pass
• Multimodal input — Unified processing of text and images

---

Core Capabilities

01 Frontend & Interactive Development

Qwen3.6 Max's progress in frontend development is one of the most obvious improvements in this generation. From design intent to runnable UI, it produces well-structured, well-styled code, and the "see results as you code" development experience is significantly improved.

• Full implementation of components and pages
• Handling of styling and interaction details
• From description to runnable UI

02 3D Scenes & Games

Outstanding at complex tasks such as 3D scene construction and game logic implementation — tasks that simultaneously test spatial reasoning, state management, and code organization skills.

• 3D scene setup and rendering logic
• Game state machines and interaction logic
• Physics and animation implementation

03 Repository-Level Problem Solving

Excellent on benchmarks like SWE-bench that require understanding an entire repository, locating problems, writing fixes, and verifying them — well suited as the core model for coding agents.

• Cross-file problem localization
• Fix generation and verification
• Structural understanding of large projects

04 Efficiency Advantages of the Hybrid Architecture

The combination of linear attention and sparse MoE routing allows the model to maintain large-parameter capabilities while controlling inference costs, with the efficiency advantage especially evident in long-context scenarios.

• Inference efficiency in long contexts
• Sparse activation controls costs
• High performance with scalability

---

Best Use Cases

Use Case	Description
Frontend development	Full implementation of components, pages, and interactions
3D & game development	Scene construction, game logic, and animation implementation
Repository-level fixes	Real-world engineering tasks like SWE-bench
Rapid prototyping	Exploratory development in the "vibe coding" style
Long-context analysis	Understanding large projects within a 1M token window
Multimodal development	UI implementation combined with design mockup images

---

Qwen3.6 Max vs. Qwen3.6 Plus vs. Qwen3.7 Max

Capability	Qwen3.6 Max Preview	Qwen3.6 Plus	Qwen3.7 Max
Model ID	qwen3.6-max-preview	qwen3.6-plus	qwen3.7-max
Positioning	Flagship of the 3.6 series	Balanced tier of the 3.6 series	Next-generation flagship
Architecture	Hybrid (linear attention + MoE)	Hybrid	Trillion-parameter MoE
Context window	1M tokens	1M tokens	1M tokens
Max output	65.5K tokens	65.5K tokens	128K tokens
Input modalities	Text, image	Text, image	Text
Focus	Frontend, 3D, repository-level problems	Balanced coding and reasoning	Agentic coding and autonomous execution

For specific billing details, refer to the real-time price card at the top of the page.

---

How to Use the Qwen3.6 Max Preview API

1. Create a SeaWhale AI API key Generate a key in the console and top up your balance.

2. You can directly provide design mockups for frontend tasks Qwen3.6 Max supports image input. Passing in design mockups or reference screenshots along with the request yields much higher fidelity than text-only descriptions.

3. Call the API

curl -X POST https://api.atalk-ai.com/v1/chat/completions \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "qwen3.6-max-preview",
    "messages": [
      {"role": "user", "content": "Use Three.js to implement an interactive 3D product showcase page with rotation and zoom support."}
    ],
    "stream": true
  }'

---

FAQ

What does hybrid architecture mean? It refers to the combination of linear attention mechanisms and sparse mixture-of-experts (MoE) routing. Linear attention reduces the computational complexity of long contexts, while sparse MoE controls activation costs while maintaining a large parameter count.

What improvements were made over the 3.5 series? Agentic coding, frontend development, and overall reasoning have all improved significantly, with the most noticeable gains in the "vibe coding" (write as you go, rapid iteration) experience.

What specific tasks is it best at? 3D scenes, game development, and repository-level problem solving are officially highlighted strengths, with standout performance on SWE-bench-style benchmarks.

How do I choose between it and Qwen3.7 Max? Qwen3.7 Max is the next-generation flagship with a higher output limit (128K vs 65.5K) and stronger agentic coding; Qwen3.6 Max supports image input, with frontend and 3D scenes as its strengths.

What does Preview mean? It means this is a preview version, and capabilities and API may be adjusted as upstream updates roll in.

What are the context and output limits? 1M token context, with a maximum output of 65,500 tokens.

---

Why Use Qwen3.6 Max Preview API with SeaWhale AI

• No Alibaba Cloud account required — Call directly with one API key
• OpenAI-compatible API — Integrate by changing two lines in your existing code
• Multiple versions coexist — Compare and switch between the Qwen3.6 and 3.7 series in the same account
• RMB settlement — Domestic models billed in RMB

API

API integration

Model IDUse this value as the model in inference requests

qwen3.6-max-preview

API KeyBearer token used to authenticate inference requests

Manage API keys

Base URLOpenAI compatible · /chat/completions

https://api.atalk-ai.com/v2/

provider OptionalSelects a service channel; omit it and the system picks the default

"provider": { "channel": "direct" }

qwen3.6-max-preview usage examples

SeaWhale AI is compatible with the OpenAI API protocol, so you can call it with the OpenAI SDK or plain HTTP requests. Streaming is enabled by default.

About the provider parameter (optional, a SeaWhale AI extension): most models are served over several channels that differ slightly in price and reliability. Add a provider field to the request body to pick one; omit it and the system selects the default channel — normal calls are unaffected.

provider is not part of the official OpenAI protocol — it is a SeaWhale AI extension that only takes effect on this platform. The OpenAI SDK allows custom fields like this to pass through; see the examples below.

Value	Channel	Best for
direct	Direct	The official upstream link, for native behavior and the full context window
stable	Preferred	Balanced availability and speed — a good default for production traffic
economical	Economy	Cost first, well suited to batch processing and price-sensitive workloads

Available channels and their prices are listed under "Pricing" above (channels vary by model). Additional notes:

• Syntax: "provider": { "channel": "direct" }.
• If the requested channel is not enabled for that model, the request automatically falls back to the default channel and returns normally — it does not error.
• With the official SDKs: in Python pass it via extra_body; in Node.js put it directly on the request object and it passes through. In TypeScript projects, add a // @ts-expect-error line to skip the type check.

cURL

curl https://api.atalk-ai.com/v2/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <API_KEY>" \
  -d '{
    "model": "qwen3.6-max-preview",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello!"}
    ],
    "provider": { "channel": "direct" },
    "stream": true
  }'
# provider is optional — remove this line to use the default channel

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.atalk-ai.com/v2",
    api_key="<API_KEY>",
)

stream = client.chat.completions.create(
    model="qwen3.6-max-preview",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"},
    ],
    stream=True,
    # Optional: pick a service channel; omit to use the default
    extra_body={"provider": {"channel": "direct"}},
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

Node.js

import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://api.atalk-ai.com/v2',
  apiKey: '<API_KEY>',
})

const stream = await client.chat.completions.create({
  model: 'qwen3.6-max-preview',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'Hello!' },
  ],
  stream: true,
  // Optional: pick a service channel; omit to use the default
  // @ts-expect-error provider is a SeaWhale AI extension, not in the OpenAI SDK types
  provider: { channel: 'direct' },
})

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? '')
}