Stop Writing 'Think Step by Step': How to Rewrite Prompts for Reasoning Models
The 2026 cognitive gap: reasoning models like Opus, GPT-5, and Gemini Deep Think already think internally, so 2023-era incantations like 'you are a top expert,' 'take a deep breath,' and 'let's think step by step' are ineffective on them — or even counterproductive. This prompt has the AI rewrite your old prompts into a reasoning-model version, with a before/after comparison table, and explains which models count as reasoning models and whether free tiers support them.
你是熟悉 2026 推理模型（如 Claude Opus、GPT-5 系列、Gemini 2.5 Deep Think）prompt 寫法的提示工程顧問。這類模型已內建思考鏈，2023-2025 那套咒語在它們身上多半無效甚至扣分。請把我的舊 prompt 改成推理模型版本。 【我的舊 prompt】 [舊prompt] 【我真正想要的結果】[目標結果] 【我打算用的模型】[使用模型] 請按以下步驟處理： 一、診斷：指出我舊 prompt 裡哪些是「過時咒語」——例如「Let's think step by step」「深呼吸」「你是世界頂尖的○○專家」這類角色灌水、要求它逐步思考的指令。簡短說明在推理模型上為什麼這些沒用或反效果（它本來就會想、你叫它想反而限制路徑）。 二、重寫：給出改寫後的新 prompt。原則是——把力氣花在「把任務講清楚」而不是「催它努力」：明確的目標、必要的脈絡與限制、你要的輸出格式與成功標準，刪掉所有戲劇性角色設定與催眠句。一句精準的任務描述勝過落落長的角色扮演。 三、新舊對照表：用表格列出「我刪掉了什麼／為什麼」，每列一項，讓我看懂差在哪。 四、使用提醒：根據我填的【使用模型】，告訴我這支算不算推理模型、免費版有沒有推理能力、什麼時候該手動開「思考／reasoning」模式、什麼任務其實用一般模型更快更省。 全程台灣正體中文。如果我的舊 prompt 其實沒什麼過時問題，就老實說「這條不太需要改」，不要硬改。
Swap the variables inside the [ ] brackets for your own content, then paste into ChatGPT / Claude / Gemini.
See what this prompt actually produces without leaving the site (live AI run, 1 credit).
This is the biggest cognitive gap right now, one that has made older tutorials almost entirely obsolete: multiple 2026 guides agree that reasoning models already think internally, and the 2023 playbook — think step by step, you are an expert, take a deep breath — is ineffective on them and can even backfire, leaving a lot of people asking 'why did my old prompt stop working?' This is even more pronounced in Traditional Chinese content — 99% of Chinese-language blogs are still teaching that 2023 playbook, so this prompt is built explicitly as a 'before/after' comparison to call out that outdated habit. Since Taiwanese readers mostly use the free tiers of ChatGPT or Gemini, step four specifically adds guidance on which models count as reasoning models, whether a free tier supports them, and when you should turn on thinking mode. Honest notes: (1) not every task benefits from a reasoning model — simple rewrites or translations are faster and cheaper on a standard model; (2) it can occasionally overcorrect and strip out context you actually needed, treating it as if it were an incantation — that's why a safeguard is built in allowing it to say 'no changes needed' if there's nothing to fix. Always fill in the model variable — the recommendations differ a lot between free/paid and reasoning/standard models.
[舊prompt]你現在在用、想升級的那段 prompt、整段貼上。例：「請你扮演世界頂尖的行銷專家，深呼吸，一步一步思考，幫我寫…」
[目標結果]你其實想拿到什麼，讓它知道改寫方向。例：「一份能直接給老闆看的競品分析，要有結論和數據」
[使用模型]你打算用哪支、付費還免費。例：「ChatGPT 免費版」「Claude Opus 付費」「Gemini 2.5 Pro」
填下面的欄位，上方 prompt 會即時替換 [方括號] 內容。填好後按「複製組好的 prompt」直接丟進工具。
你是熟悉 2026 推理模型（如 Claude Opus、GPT-5 系列、Gemini 2.5 Deep Think）prompt 寫法的提示工程顧問。這類模型已內建思考鏈，2023-2025 那套咒語在它們身上多半無效甚至扣分。請把我的舊 prompt 改成推理模型版本。 【我的舊 prompt】 [舊prompt] 【我真正想要的結果】[目標結果] 【我打算用的模型】[使用模型] 請按以下步驟處理： 一、診斷：指出我舊 prompt 裡哪些是「過時咒語」——例如「Let's think step by step」「深呼吸」「你是世界頂尖的○○專家」這類角色灌水、要求它逐步思考的指令。簡短說明在推理模型上為什麼這些沒用或反效果（它本來就會想、你叫它想反而限制路徑）。 二、重寫：給出改寫後的新 prompt。原則是——把力氣花在「把任務講清楚」而不是「催它努力」：明確的目標、必要的脈絡與限制、你要的輸出格式與成功標準，刪掉所有戲劇性角色設定與催眠句。一句精準的任務描述勝過落落長的角色扮演。 三、新舊對照表：用表格列出「我刪掉了什麼／為什麼」，每列一項，讓我看懂差在哪。 四、使用提醒：根據我填的【使用模型】，告訴我這支算不算推理模型、免費版有沒有推理能力、什麼時候該手動開「思考／reasoning」模式、什麼任務其實用一般模型更快更省。 全程台灣正體中文。如果我的舊 prompt 其實沒什麼過時問題，就老實說「這條不太需要改」，不要硬改。
Suno Engineer's Mindset: 4 Steps to a Song That Doesn't Sound Like AI
A studio engineer's breakdown of Suno's fatal weaknesses (fried vocals, high-frequency artifacts), plus a 4-step DAW workflow and a Suno Studio cleanup prompt.
5 Claude Weekly Workflows That Stuck After 6 Months
Proposal generator / meeting processor / content repurposer / Friday review / shutdown reset — out of 40 I tried, only these 5 survived, each saving 30+ minutes per run.