Prompts for Qwen: clear sections, delimiters and one example

How to prompt Qwen3.8 Max, Qwen3.7 Plus and Qwen3.8 Flash on Alibaba Cloud Model Studio: the section framework Alibaba recommends, ### delimiters, examples and JSON mode.

Last updated: · Markdown version

What changes with this model

Even if prompt structure is similar, each model has biases: some follow strict instructions better, some are more creative, and some need examples. These templates are designed to reduce ambiguity and improve consistency.

Recommendation: define output format (ideally with a checklist), add constraints, and if you need precision, use few-shot or require strict JSON.

Current models Promptea supports

Data taken from each provider’s official documentation; every entry shows the date it was last verified.

Qwen3.8 Max

default

Alibaba Cloud Model Studio's most capable Qwen (Qwen3.8, Aug 2026): 1M context, text, image and video input, thinking on by default and built-in web search (enable_search).

Context window
1M tokens
Max output
131K tokens
Input / output
text, image, video / text
Latency
slower (deliberate)
Structured output
yes
Tool use / native search
yes / yes

How to prompt it: Thinking is on by default (enable_thinking, with thinking_budget to cap it); turn it off for quick answers. For JSON use response_format json_schema with strict, or json_object with the word 'JSON' in the prompt (otherwise the API returns an error).

Verified on November 14, 2026 · official source

Qwen3.7 Plus

Alibaba's recommended starting point: balances capability and cost. 1M context; text, image and video input.

Context window
1M tokens
Max output
131K tokens
Input / output
text, image, video / text
Latency
balanced
Structured output
yes
Tool use / native search
yes / no

How to prompt it: Thinking is on by default (enable_thinking, with thinking_budget to cap it); turn it off for quick answers. For JSON use response_format json_schema with strict, or json_object with the word 'JSON' in the prompt (otherwise the API returns an error).

Verified on November 14, 2026 · official source

Qwen3.8 Flash

The fast, low-cost Qwen (Qwen3.8 Flash, Aug 2026), with the same 1M context. For short tasks and high volume.

Context window
1M tokens
Max output
131K tokens
Input / output
text, image, video / text
Latency
fast
Structured output
yes
Tool use / native search
yes / no

How to prompt it: Thinking is on by default (enable_thinking, with thinking_budget to cap it); turn it off for quick answers. For JSON use response_format json_schema with strict, or json_object with the word 'JSON' in the prompt (otherwise the API returns an error).

Verified on November 14, 2026 · official source

Quick tips
Templates
Six-part brief
### Context
[background the model needs]
### Objective
[the one task to complete]
### Style
[e.g. a product manager's status note]
### Tone
[e.g. plain, direct]
### Audience
[who reads it and what they already know]
### Response
[exact format: e.g. 5 bullets, max 15 words each]
Opens the analyzer with the prompt prefilled.
Open in Promptea
JSON extraction (json_object mode)
Extract the fields below from the text and return JSON only.
Fields: "name" (string), "company" (string or null), "email" (string or null).
If a field is missing, use null; do not guess.
Example: {"name": "Ana Ruiz", "company": "Tasklight", "email": null}
###
[paste the text here]
###
Opens the analyzer with the prompt prefilled.
Open in Promptea
FAQ
Which Qwen model should I pick?
Alibaba calls Qwen3.7 Plus the recommended starting point (capability and cost in balance). Use Qwen3.8 Max for the hardest reasoning and long multimodal inputs, and Qwen3.8 Flash for short, high-volume tasks.
Should I turn thinking off?
Thinking is on by default. Turn it off (enable_thinking: false) for quick, simple answers, or cap it with thinking_budget; keep it on for reasoning, math and code.