节奏 · 档位、模型、成本与同类分工Cadence · Plans, Models, Cost, and Rivals.
选档不是看价格,是看你撞哪堵墙:额度、模型,还是速度。这一章把 Plus、Pro(5× / 20× 两档)、API key 旁路摆开,讲清 GPT-5.6 三兄弟的分工 —— Sol 管复杂开放的活,Terra 是日常主力,Luna 跑量大清晰的活 —— 以及 credits 怎么算、fast mode 什么时候值 2.5 倍的价、官方那五条省额度纪律为什么条条都在讲上下文。再划同类分工线:Claude Code、Cursor、Copilot、Warp 各自的主场,已经付着 ChatGPT 的人什么时候该加第二份订阅 —— 以及为什么多数时候不该。Choosing a plan isn't about price — it's about which wall you hit: quota, model, or speed. This chapter lays out Plus, Pro (its 5x and 20x tiers), and the API-key side road; splits the GPT-5.6 trio — Sol for complex open-ended work, Terra the everyday workhorse, Luna for clear high-volume jobs; and covers how credits are counted, when fast mode earns its 2.5x rate, and why every one of the five official quota-saving tips is really about context. Then the rivalry lines: where Claude Code, Cursor, Copilot, and Warp each win, when someone already paying for ChatGPT should add a second subscription — and why most of the time they shouldn't.
11 分钟 · 初稿 2026.0711 Min · Drafted 2026.07
选档的错误问法是「哪档划算」,正确问法是「我撞的是哪堵墙」。额度不够,是一堵墙;默认模型不够强,是另一堵;等不起响应速度,是第三堵 —— 三堵墙对应三个不同的开关,只有一个真需要花更多钱。这一章把档位、模型、力度、成本四个旋钮摆开,最后划一条和同类的分工线:已经付着 ChatGPT 的人,什么时候该加第二份订阅 —— 以及为什么多数时候不该。The wrong question for picking a plan is 'which tier is the best deal'; the right one is 'which wall am I hitting.' Not enough quota is one wall; a default model that's not strong enough is another; responses too slow to wait for is a third — three walls, three different switches, and only one actually costs more money. This chapter lays out the four dials — plan, model, effort, cost — then draws the line of division with the rivals: when someone already paying for ChatGPT should add a second subscription, and why most of the time they shouldn't.
— I
档位与三堵墙Plans, and the Three Walls.
先把价目摆平:Free $0、Go $8、Plus $20、Pro 从 $100 起(按比 Plus 多 5× 还是 20× 用量分两档)、Business $20/人·月、Enterprise。另有一条 API key 旁路,按 token 计费、不含云端功能(GitHub 审查、Slack 那些)。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.要紧的一条:Codex 和 ChatGPT Work 共用同一个 usage 池 —— 你在 Work 里跑掉的和在 Codex 里跑掉的,是同一笔额度。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.Flatten the price list first: Free $0, Go $8, Plus $20, Pro from $100 (in two tiers by whether you want 5x or 20x more usage than Plus), Business $20/user·mo, Enterprise. There's also an API-key side road, billed per token and excluding cloud features (GitHub review, Slack, and such).1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. One thing that matters: Codex and ChatGPT Work share one usage pool — what you burn in Work and what you burn in Codex draw on the same allowance.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.
你撞的墙The wall you hit
症状Symptom
开关The switch
额度Quota
5 小时窗口内就见底,活排着队Dry within the 5-hour window, work queuing
升 Pro(5× / 20×),或先做第 III 节的省额度纪律Move to Pro (5x / 20x) — or apply section III's discipline first
模型Model
难题上明显吃力Visibly straining on hard problems
换 Sol、升思考力度(第 II 节)Switch to Sol, raise reasoning effort (section II)
速度Speed
交互时等不起Interactive waits you can't afford
/fast(走 credits,另一本账)/fast (credits — a separate ledger)
额度墙的形状值得看清:本地消息按模型和任务大小,每 5 小时给一个范围 —— Plus 下 Sol 大概 15–90 条、Terra 20–110、Luna 50–280(任务越大越复杂,吃得越多);Pro 的 5× / 20× 就是把这个范围整体抬 5 倍、20 倍。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.还有一条容易忽略的:本地和云端共享同一个 5 小时窗口 —— 你在云端并行派一批,吃的和本地是同一格额度。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.判断顺序照表走:先确认撞的是哪堵墙,再动钱包 —— 三堵墙里只有额度墙必须用钱解决。The quota wall's shape is worth seeing: local messages, by model and task size, give a range per 5 hours — on Plus roughly Sol 15–90, Terra 20–110, Luna 50–280 (bigger, more complex tasks eat more); Pro's 5x / 20x lifts that whole range 5x, 20x.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. And one easily missed: local and cloud share the same 5-hour window — dispatch a batch in parallel to the cloud and it draws on the same allowance as local.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. Follow the table's order: identify the wall first, then reach for the wallet — of the three, only the quota wall must be solved with money.
— II
三兄弟与力度:两个旋钮The Trio and the Effort Dial: Two Knobs.
模型的日常分工是三个 GPT-5.6 兄弟。Sol 管细节与打磨 —— 复杂、开放、高价值的活(旗舰,拿不准就用它);Terra 是务实全能的日常主力,原先给 GPT-5.5 的活自然落到它;Luna 跑清晰可重复、量大的活:抽取、分类、转换。2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17.默认的 Power 设置就是 Sol 加 medium 推理。2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17.最顺手的用法不是主对话换来换去,而是第 7 章那招:难活主力用 Sol,批量的读 / 扫 / 整理派给 Terra 或 Luna 的并行任务。The daily division of models is three GPT-5.6 siblings. Sol for detail and polish — complex, open-ended, high-value work (the flagship; use it when unsure); Terra the pragmatic all-rounder and everyday workhorse, where work once given GPT-5.5 naturally lands; Luna for clear, repeatable, high-volume work: extraction, classification, transformation.2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17. The default Power setting is Sol plus medium reasoning.2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17. The smoothest arrangement isn't swapping your main conversation around but chapter 7's move: Sol for the hard work, batched read / scan / tidy jobs to parallel tasks on Terra or Luna.第二个旋钮是推理力度:从 Light(应用)/ Low(CLI)、Medium、到 High、Extra High,控制它每一步想多深 —— 难题升档、机械活降档,同一个模型,力度对了常比换模型便宜。2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17.再上面有两个特殊档:Max 给单个任务更多推理时间,啃最硬的问题;Ultra 用 subagents 并行拆解,把一件能分块的复杂活铺开来跑 —— 多数活两个都用不上。2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17.还有一个 Pro 专属的速度型模型 Codex-Spark(研究预览):near-instant 的纯文本迭代,和 fast mode 不是一回事,它是独立模型、有自己的限额。3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17.The second knob is reasoning effort: from Light (app) / Low (CLI), Medium, up to High, Extra High, controlling how deeply it thinks per step — raise it for hard problems, drop it for mechanical ones; same model, right effort, often cheaper than switching models.2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17. Above that are two special settings: Max gives one task more reasoning time for the hardest problems; Ultra uses subagents to split a divisible complex job in parallel — most work needs neither.2注 2Note 2OpenAI Codex 文档 · Models —— Codex 有三个 GPT-5.6 模型:Sol(细节与打磨,复杂开放 / 高价值任务的旗舰)、Terra(务实全能,日常主力,原先给 GPT-5.5 的活的自然起点)、Luna(清晰可重复、量大的活:抽取、分类、转换)。默认 Power 设置 = Sol + medium 推理。推理力度:Light(应用)/ Low(CLI)、Medium、High、Extra High;Max 给单任务更多推理时间,Ultra 用 subagents 并行拆解复杂任务。gpt-5.2 与 gpt-5.3-codex 对 ChatGPT 登录已弃用。云端对话默认模型不可改。截至 2026-07-17。OpenAI Codex Docs · Models — Codex has three GPT-5.6 models: Sol (detail and polish, the flagship for complex / open-ended, high-value tasks), Terra (pragmatic all-rounder, the everyday workhorse, a natural start for work previously given GPT-5.5), Luna (clear, repeatable, high-volume work: extraction, classification, transformation). The default Power setting = Sol + medium reasoning. Reasoning effort: Light (app) / Low (CLI), Medium, High, Extra High; Max gives a single task more reasoning time, Ultra uses subagents to split a complex task in parallel. gpt-5.2 and gpt-5.3-codex are deprecated for ChatGPT sign-in. The cloud-chat default model can't be changed. As of 2026-07-17. And there's a Pro-only speed model, Codex-Spark (research preview): near-instant text-only iteration, not the same as fast mode — it's a separate model with its own limits.3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17.
— III
成本:credits、fast,和五条纪律Cost: Credits, Fast, and Five Disciplines.
撞了额度墙,Codex 让当前这轮先跑完(公平使用范围内),然后 Plus / Pro 可以买 credits 续用,不必升档。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.credits 是把 token 折成的统一单位,按模型算 —— 大致上 Luna 最省、Terra 居中、Sol 最贵,这也和三兄弟的分工对得上:量大的活交便宜模型,是省钱也是省额度。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.Hit the quota wall and Codex lets the current turn finish (within fair use), after which Plus / Pro can buy credits to continue without upgrading.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. Credits are a unified unit that folds tokens together, counted by model — roughly Luna cheapest, Terra middle, Sol priciest, which matches the trio's division: sending high-volume work to the cheaper model saves both money and quota.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.Fast mode 是用钱换等待,想清楚再开:它把支持的模型提速 1.5×,代价是按更高倍率吃 credits —— GPT-5.6 / 5.5 是标准的 2.5×,GPT-5.4 是 2×;/fast on 切。3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17.注意它是 credit 特性、只有 ChatGPT 登录能用;API key 走的是 API token 价、没有这个倍率。3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17.正确用法窄而清晰:赶工的交互时段开(快速迭代、现场调试),长的自动化任务与 CI 一律关。Fast mode trades money for waiting — decide before you flip it: it speeds a supported model up 1.5x, at the cost of consuming credits at a higher multiplier — 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on.3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17. Note it's a credit feature usable only with ChatGPT sign-in; an API key uses API token pricing with no such multiplier.3注 3Note 3OpenAI Codex 文档 · Speed —— Fast mode 把支持的模型提速 1.5×,按更高倍率吃 credits:GPT-5.6 / 5.5 为标准的 2.5×,GPT-5.4 为 2×;/fast on|off|status 切,或 config 里 service_tier = "fast"。仅 ChatGPT 登录可用(是 credit 特性);API key 走 API token 价、无 credit 倍率。GPT-5.3-Codex-Spark 是另一个更快、能力较弱的独立模型(研究预览,仅 ChatGPT Pro),有自己的用量限额,和 fast mode 不是一回事。截至 2026-07-17。OpenAI Codex Docs · Speed — Fast mode speeds a supported model up by 1.5x, consuming credits at a higher multiplier: 2.5x the standard rate for GPT-5.6 / 5.5, 2x for GPT-5.4; toggle with /fast on|off|status or service_tier = "fast" in config. Available only with ChatGPT sign-in (it's a credit feature); an API key uses API token pricing with no credit multiplier. GPT-5.3-Codex-Spark is a separate faster, less-capable model (research preview, ChatGPT Pro only) with its own limits, distinct from fast mode. As of 2026-07-17. Its correct use is narrow and clear: on for interactive crunch (rapid iteration, live debugging), off for long automated runs and CI.最省钱的其实不是任何一个开关,是纪律。官方那五条省额度建议,条条都在讲同一件事 —— 管住上下文:prompt 写精、只喂相关文件、把 AGENTS.md 拆小(第 5 章)、少接 MCP(第 6 章)、常规活降到便宜模型。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.前面几章的每个习惯,在账单上都有回声:上下文纪律,就是额度纪律。What saves most isn't any switch — it's discipline. The five official quota-saving tips all say the same thing: manage context — tighten prompts, feed only relevant files, split AGENTS.md small (chapter 5), wire in fewer MCP servers (chapter 6), drop routine work to the cheaper model.1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. Every habit from the earlier chapters echoes on the bill: context discipline is quota discipline.
— IV
同类分工:入口跟着活走The Rivals: Entrances Follow the Work.
最后划分工线 —— 不是黑稿,是各自的主场。Claude Code(Pro $20 / Max 从 $100,均含在订阅里)的主场在终端里的编排:计划、记忆、subagents、hooks 拼成的那套「派活 + 验收」循环,是它写得最透的地方。4注 4Note 4Claude · Pricing —— Pro:月付 $20,年付折合每月 $17(一次预付 $200);Max:从 $100 起,可选比 Pro 多 5× 或 20× 用量两档;Pro 与 Max 均含 Claude Code。截至 2026-07-17。Claude · Pricing — Pro: $20 billed monthly, or $17/month on annual billing ($200 up front); Max: from $100, with a choice of 5x or 20x more usage than Pro; both Pro and Max include Claude Code. As of 2026-07-17.Cursor(Pro $20 起,顶上还有 Pro+ / Ultra)的主场在编辑器:你要「陪着改、盯每一步 diff」的工作节奏时,它的 agent 工作台仍是最顺的入口。5注 5Note 5Cursor · Pricing —— Hobby 免费;个人付费从 Pro $20/月起(页面另有 Pro+ / Ultra 更高档);Teams $40/座·月,Enterprise 定制。卖点为 Agent 请求、code review、cloud agents。截至 2026-07-17。Cursor · Pricing — Hobby free; paid individual from Pro at $20/mo (the page also has higher Pro+ / Ultra tiers); Teams $40/seat·mo, Enterprise custom. The pitch is Agent requests, code review, and cloud agents. As of 2026-07-17.Copilot(Pro $10 起)的主场在 GitHub 本体:issue 直接派、收回一个 PR,甚至能把活转派给 Claude、Codex 这些第三方 agent —— 你的团队整个活在 GitHub 流程里时,它是那条流程自带的手。6注 6Note 6GitHub Copilot · Plans —— Free / Pro $10 / Pro+ $39 / Max $100(每人·月),另有 Business 与 Enterprise;卖点原话「把活派给 Copilot,它创建一个 pull request」,且 Pro+ / Max 可把任务转派给 Claude、OpenAI Codex 等第三方编码 agent。截至 2026-07-17。GitHub Copilot · Plans — Free / Pro $10 / Pro+ $39 / Max $100 (per user·mo), plus Business and Enterprise; the pitch verbatim 'Assign work to Copilot and it creates a pull request,' with Pro+ / Max able to delegate tasks to third-party coding agents like Claude and OpenAI Codex. As of 2026-07-17.Warp 自我定位「面向 agentic coding 的现代终端」(Build $20/月):活以终端会话本身为中心时,它贴得更近。7注 7Note 7Warp · Pricing —— 自我定位「面向 agentic coding 的现代终端」;Free / Build $20 / Max $200(月付),另有 Business $50/人·月(年付 $45)与 Enterprise。截至 2026-07-17。Warp · Pricing — self-positioned as 'a modern terminal for agentic coding'; Free / Build $20 / Max $200 monthly, plus Business $50/user·mo ($45 annual) and Enterprise. As of 2026-07-17.Now the division — no hit piece, just home turf. Claude Code (Pro $20 / Max from $100, both bundled in the subscription) owns orchestration in the terminal: the delegate-and-verify loop built from plan, memory, subagents, and hooks is where it's written most thoroughly.4注 4Note 4Claude · Pricing —— Pro:月付 $20,年付折合每月 $17(一次预付 $200);Max:从 $100 起,可选比 Pro 多 5× 或 20× 用量两档;Pro 与 Max 均含 Claude Code。截至 2026-07-17。Claude · Pricing — Pro: $20 billed monthly, or $17/month on annual billing ($200 up front); Max: from $100, with a choice of 5x or 20x more usage than Pro; both Pro and Max include Claude Code. As of 2026-07-17.Cursor (from Pro $20, with Pro+ / Ultra above) owns the editor: when your rhythm is 'work alongside, watch every diff,' its agent workbench remains the smoothest entrance.5注 5Note 5Cursor · Pricing —— Hobby 免费;个人付费从 Pro $20/月起(页面另有 Pro+ / Ultra 更高档);Teams $40/座·月,Enterprise 定制。卖点为 Agent 请求、code review、cloud agents。截至 2026-07-17。Cursor · Pricing — Hobby free; paid individual from Pro at $20/mo (the page also has higher Pro+ / Ultra tiers); Teams $40/seat·mo, Enterprise custom. The pitch is Agent requests, code review, and cloud agents. As of 2026-07-17.Copilot (Pro from $10) owns GitHub itself: assign an issue, collect a PR — it can even delegate onward to third-party agents like Claude and Codex — so when your team lives entirely in GitHub flow, it's the hand that flow comes with.6注 6Note 6GitHub Copilot · Plans —— Free / Pro $10 / Pro+ $39 / Max $100(每人·月),另有 Business 与 Enterprise;卖点原话「把活派给 Copilot,它创建一个 pull request」,且 Pro+ / Max 可把任务转派给 Claude、OpenAI Codex 等第三方编码 agent。截至 2026-07-17。GitHub Copilot · Plans — Free / Pro $10 / Pro+ $39 / Max $100 (per user·mo), plus Business and Enterprise; the pitch verbatim 'Assign work to Copilot and it creates a pull request,' with Pro+ / Max able to delegate tasks to third-party coding agents like Claude and OpenAI Codex. As of 2026-07-17.Warp positions itself as 'a modern terminal for agentic coding' (Build $20/mo): when work centers on the terminal session itself, it sits closer.7注 7Note 7Warp · Pricing —— 自我定位「面向 agentic coding 的现代终端」;Free / Build $20 / Max $200(月付),另有 Business $50/人·月(年付 $45)与 Enterprise。截至 2026-07-17。Warp · Pricing — self-positioned as 'a modern terminal for agentic coding'; Free / Build $20 / Max $200 monthly, plus Business $50/user·mo ($45 annual) and Enterprise. As of 2026-07-17.Codex 自己的主场,是这本书的整条主线:四面开工 + 云端异步 —— 已经付着 ChatGPT、又常做「派一批上云、回来收 PR」的人,它是那个不用额外付费就能用满的选项(和 Work 共用额度池)。1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17.组合的原则一句话:入口跟着活走,订阅别叠着买 —— 每加一份订阅,都该对应一类你真在做、且 Codex 干得不如它的活;先把手里这份 ChatGPT 用满,再谈第二份。Codex's own home turf is this book's whole through-line: four fronts plus async cloud — for anyone already paying for ChatGPT who often does 'send a batch to the cloud, collect PRs back,' it's the option you can use to the full at no extra cost (sharing the pool with Work).1注 1Note 1OpenAI Codex 文档 · Pricing —— 档位:Free $0、Go $8、Plus $20、Pro 从 $100(选比 Plus 多 5× 或 20× 用量)、Business $20/人·月(年付,月付 $25)、Enterprise & Edu;API key 旁路按 token 计费、不含云端功能。Codex 与 ChatGPT Work 共用同一 usage 池。本地消息按模型和任务大小每 5 小时给一个范围:Plus 下 Sol 15–90、Terra 20–110、Luna 50–280;Pro 5× 约为 5 倍,Pro 20× 约为 20 倍;本地与云端共享同一个 5 小时窗口。撞额度时活能跑完当前轮,Plus/Pro 可买 credits 续用。截至 2026-07-17。OpenAI Codex Docs · Pricing — tiers: Free $0, Go $8, Plus $20, Pro from $100 (choose 5x or 20x more usage than Plus), Business $20/user·mo (annual; $25 monthly), Enterprise & Edu; the API-key side road bills per token and excludes cloud features. Codex shares one usage pool with ChatGPT Work. Local messages per 5 hours give a range by model and task size: on Plus, Sol 15–90, Terra 20–110, Luna 50–280; Pro 5x is roughly 5x, Pro 20x roughly 20x; local and cloud share one 5-hour window. On hitting a limit the current turn finishes, and Plus/Pro can buy credits to continue. As of 2026-07-17. The combination principle in one line: entrances follow the work; don't stack subscriptions — each added one should map to a kind of work you actually do and that Codex does worse; use the ChatGPT you already have to the full before discussing a second.动手 · 用一周数据代替直觉:Hands-on · replace instinct with a week of data:
记录一周你每次「用不了」的原因:额度见底、模型吃力、还是嫌慢 —— 月底对着三堵墙的表决定动不动钱包。Log every can't-use moment for a week: quota dry, model straining, or too slow — then decide against the three-wall table whether the wallet moves.
把批量的读 / 扫活降到 Luna、难活留给 Sol,跑一周看额度曲线的变化。Downshift batched read / scan work to Luna while hard work stays on Sol; run a week and watch the quota curve.
只在赶工的那一小时开 /fast,完事就关 —— 月底看 credits 那本账,判断这个「快」值多少钱。Flip /fast on only for the crunch hour and off after — then read the credits ledger at month's end to price what 'fast' was worth.
入口跟着活走,
订阅别叠着买.
Entrances follow the work;
don't stack subscriptions.
Aklman Library
— 讨论Discussion
讨论Discussion.
评论区初始化中…Initializing comments…
01 / 01
没有匹配结果No matches.
换个关键词，或按 Esc 回到页面Try another keyword, or press Esc to return