抄作业 / 做成了什么 / Grok Bot

做了什么

Staff (kevinn, 2026-09-23): there is no per-agent reasoning-effort control in Bot or chat settings—Grok Bot chooses how much reasoning a turn gets, and you cannot set it per bot, role, or template yet (team aware).

场景

未归类

资料完整度

• ✓ 作者与原链接有
• ✓ 做了什么有
• ✓ 来源原文摘录有
• — 截图 / 媒体来源未提供
• — 步骤 / 章节来源未提供
• — 提示词 / 配置原文来源未提供
• — 结果 / 数字来源未提供
• ✓ 风险提示有
• ✓ 相关条目有

— 表示来源没有给出，本站不补写。

来源原文摘录

Feature request for product/service – Other – Describe the request Please add a per-agent “reasoning effort” control in Bot / chat settings — for example low / mid / full, or a way to dial down (or skip) deep reasoning on agents that do not need it. Why In a multi-agent setup, not every seat should run the full reasoning stack. Some roles are fast replies, mid-level conversation, or light coordination. Others need depth. Today it feels like everyone gets the heavy path by default, which slows group / table chats and burns weekly usage on seats that only needed a quick turn. Voice already feels

展开全文

like a stripped-down, faster path — so the capability is clearly there. Exposing the same kind of knob for text / multi-agent chat would make multi-bot conversations snappier and more natural, and stretch the weekly allowance further. Concrete ask …

摘自论坛首帖，照录原文，未翻译。

提示词

来源没有给出可直接复制的提示词原文。

怎么复制

1. 读原文：forum.cursor.com ↗
2. 把原文里的步骤/提示词整理成一条消息交给 Bot
3. 先手动跑一次，再决定频率

风险与限制

• 定时运行会持续消耗额度：频率从低开始，跑稳再加
• 作者的结果不保证能复现，以原帖为准

用量预估

用金属计算器估一下