Thinking | Qwen Cloud

Thinking (reasoning) models reason before answering — outputting reasoning_content (Chat Completions / DashScope) or reasoning_summary_text events (Responses API). Models support thinking in one of two modes:

Hybrid: toggle thinking on or off per request with enable_thinking. Qwen3.5 (enabled by default), Qwen3, Qwen3-VL, Qwen3-Omni (disabled by default).
Thinking-only: always thinks — cannot be disabled. QwQ, -thinking variants.

Enable thinking

OpenAI Chat Completions
OpenAI Responses API
DashScope

import os
from openai import OpenAI
client = OpenAI(api_key=os.getenv("DASHSCOPE_API_KEY"), base_url="https://dashscope-intl.aliyuncs.com/compatible-mode/v1")

completion = client.chat.completions.create(
  model="qwen3.6-plus",
  messages=[{"role": "user", "content": "If 3x + 7 = 22, what is x?"}],
  extra_body={"enable_thinking": True},              # ← enable thinking
  stream=True,
)
for chunk in completion:
  if not chunk.choices:
    continue
  delta = chunk.choices[0].delta
  if hasattr(delta, "reasoning_content") and delta.reasoning_content:
    print(delta.reasoning_content, end="", flush=True)  # ← phase 1: thinking
  if hasattr(delta, "content") and delta.content:
    print(delta.content, end="", flush=True)             # ← phase 2: answer

Thinking arrives as response.reasoning_summary_text.delta events, followed by response.output_text.delta for the answer.

import os
from openai import OpenAI
client = OpenAI(api_key=os.getenv("DASHSCOPE_API_KEY"), base_url="https://dashscope-intl.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1")

stream = client.responses.create(
  model="qwen3.6-plus",
  input="If 3x + 7 = 22, what is x?",
  extra_body={"enable_thinking": True},              # ← enable thinking
  stream=True,
)
for chunk in stream:
  if chunk.type == "response.reasoning_summary_text.delta":
    print(chunk.delta, end="", flush=True)         # ← phase 1: thinking
  elif chunk.type == "response.output_text.delta":
    print(chunk.delta, end="", flush=True)         # ← phase 2: answer

from dashscope import MultiModalConversation

responses = MultiModalConversation.call(
  model="qwen3.6-plus",
  messages=[{"role": "user", "content": [{"text": "If 3x + 7 = 22, what is x?"}]}],
  enable_thinking=True,                              # ← enable thinking
  stream=True,
  incremental_output=True,                           # ← recommended: new tokens only
)
for chunk in responses:
  msg = chunk.output.choices[0].message
  if msg.reasoning_content:
    print(msg.reasoning_content, end="", flush=True)   # ← phase 1: thinking
  if msg.content and msg.content[0].get("text"):
    print(msg.content[0]["text"], end="", flush=True)   # ← phase 2: answer

Control thinking depth

Token budget

Use thinking_budget to cap the number of thinking tokens. If the limit is reached, the model stops thinking and generates its answer immediately. All thinking-capable models from Qwen3 onward support this parameter. Chat Completions and DashScope only — not supported by the Responses API.

OpenAI Chat Completions
DashScope

extra_body={"enable_thinking": True, "thinking_budget": 500}

Prompt-level control

With enable_thinking: true, add /no_think to skip thinking for one turn. /think restores it. Last instruction wins. Supported by open-source Qwen3 hybrid models, qwen-plus-2025-04-28, and qwen-turbo-2025-04-28.

Function calling with thinking mode

When thinking is enabled during function calling, the model reasons about which tools to call and how to use results before responding. The response includes reasoning_content before each tool call. Key points:

Pass enable_thinking: true alongside your tools array — no other config needed.
In multi-turn tool-call flows, include the assistant's reasoning_content when sending tool results back. Omitting it degrades accuracy.
Streaming delivers thinking tokens first, then tool call deltas. See Streaming with tool calls for the parse pattern.
thinking_budget works the same as in regular thinking mode.

Thinking mode is most valuable for complex tool orchestration — multi-step reasoning about which tools to call, parameter selection, and result interpretation. For simple single-tool calls, the overhead may not be worth it.

Notes

Streaming required for most models: Non-streaming is only supported by Qwen3.5 Plus/Flash, Qwen3 Max, Qwen Plus/Flash/Turbo (commercial), and all open-source models. Streaming is always recommended to avoid timeout risks.
No audio output in thinking mode (Qwen3-Omni): Text and image inputs work normally; audio output is not available when thinking is enabled.

​Enable thinking

​Control thinking depth

​Token budget

​Prompt-level control

​Function calling with thinking mode

​Notes

Enable thinking

Control thinking depth

Token budget

Prompt-level control

Function calling with thinking mode

Notes