DeerAPI Plus icon

DeerAPI Plus

AI 全能工具箱：对话、思维链、图片、视频、向量、提示词优化、虚拟试穿 | AI Toolkit: Chat, Thinking, Image, Video, Embeddings, Prompt, Virtual Try-On

Actions11

🧠 深度思考 / Thinking → 生成 / Generate

AI-generated

Summary

Generate a deep-thinking response using a thinking model with a token budget for the reasoning process.

Inputs

  • model (required) — The thinking model to use (loaded via getThinkingModels).
  • customModel — Custom model ID when model is set to '__custom'.
  • userPrompt (required) — The problem or question to think about.
  • budgetTokens — Token budget for the thinking process (1–10000). Default: 5000.
  • additionalOptions.systemPrompt — System prompt to set model behavior. Default: 'You are a helpful assistant.'.
  • additionalOptions.maxTokens — Maximum tokens in the model response (1–16000). Default: 8192.
  • additionalOptions.simplify — Whether to return a simplified output with only key fields. Default: true.
  • additionalOptions.extraBodyFields — Additional JSON fields to merge into the request body. Default: empty string.

Output shape

object

When 'simplify' is true, only key fields are returned; otherwise includes usage and raw response.

Examples

Example 1: Use default model, a user prompt, and default token budget.

model: gemini-3-flash-preview-thinking, userPrompt: 'Explain quantum entanglement in simple terms', budgetTokens: 5000

Links

Discussion