你好AI|nihao-ai.com 站点公告
SITE UPDATES · OCTOBER 2026

10月站点更新公告

本月带来 GPT-6 自动更新、Work 编程账号、GPT Go 绘图专区和绘图余量说明。

全部已生效 最后更新:10月12日 官网原生直连

10月最新公告

GPT-6 自动更新、Work 编程账号、GPT Go 绘图专区与绘图余量说明

模型更新

GPT-6 已同步,自动调用最新模型

你好AI已同步官方最新 GPT-6 模型。进入 GPT 号池后无需手动选择,系统会自动调用当前可用的最新 GPT-6 模型。

自动调用:无需手动切换,系统会自动使用最新可用的 GPT-6 模型。
直接使用:进入 GPT 号池后正常发起对话即可。
新增账号类型

新增 Work(Codex)账号,更适合编程与长任务

GPT 号池现已增加标题为「Work」的账号类型,即原 Codex 账号,使用时会消耗 Codex 额度。该类型针对编程、代码分析和长时间任务进行了优化,更适合处理复杂、连续性的工作。

开放范围:Work 账号面向号池所有会员开放,无需额外开通专属套餐,登录后即可选择使用。
适用场景:适合编程、代码分析、项目开发和需要持续处理的复杂任务。
绘图专区

新增 GPT Go 绘图专区,专为生图用户保障用量

为保障绘图用户的使用体验,号池新增 GPT Go 绘图专用区块。需要生图时,可以翻到号池下方,找到 GPT Go 区块使用。

使用方式:进入 GPT 号池后向下翻页,选择 GPT Go 绘图专用区块。
推荐使用:需要生图时建议优先使用该专区,获得更稳定的绘图体验。
使用提醒

GPT 绘图余量显示异常,请以实际使用为准

目前 GPT 官方的绘图余量接口存在异常,无法始终准确返回账号的实际余量。由于你好AI直接对接官方,官方接口异常也会影响页面上的余量显示。

显示说明:账号卡片三个小圆点中的绘图余量暂时无法作为准确的余量判断标准。
遇到生图限制怎么办:如果提示当前账号已达到生图次数限制,并不代表会员无法继续使用。请返回主页重新选择一个账号,然后点击新建对话重新提问,不要从历史记录进入原对话。更换账号并开启新对话后即可继续尝试使用。
使用建议:生图用户建议优先使用下方的 GPT Go 绘图专区。我们会持续优化账号调度,尽可能保障大家在盲选账号的情况下顺畅使用。

9月历史公告

以下为 9 月已发布的站点更新内容,保留在此供用户回看。

最新模型

GPT 官方最新生图模型 Image 2.5 已发布

GPT Image 2.5 现已加入 GPT 号池,可直接调用。发送指令生成图片时,系统默认使用 Image 2.5 模型。

直接使用:进入 GPT 对话页发送生图指令,无需额外切换模型。
默认模型:图片生成请求将默认调用 GPT Image 2.5。
新模型上线

GPT-6 模型现已加入 GPT 号池

大家期待的 GPT-6 模型,现在已经可以在 GPT 号池里使用了。GPT-6 才刚刚发布,官网目前还没有推送到 GPT Plus 网页端。为了让大家不用继续等,我们第一时间通过官方 API 接入了 GPT-6 模型,欢迎大家体验。

怎么使用:进入 GPT 使用页面后,点击对话框右侧的思考强度选项(即时 / 中 / 高),再从下拉菜单里选择 GPT-6 系列模型,就可以开始使用了。
请注意:你好AI号池一直以来都是后台直接对接官网真实 GPT Plus 订阅账号,GPT-5.6 Sol 等常规模型均按这个方式提供服务。GPT-6 目前官网还没有推送到 GPT Plus 网页端,为了让大家提前体验,因此我们通过官方 API 接入。由于两种模型的接入方式不同,请不要在同一个对话里混用两种模型,否则可能导致该对话中的部分聊天记录缺失。使用 GPT-6 时,建议单独新建一个对话
会员专属福利

月会员及以上套餐用户赠送视频生图创作站 1000 积分

所有已开通任意 AI 月会员及以上套餐的用户,均可获赠视频生图创作站 1000 积分兑换码。

领取方式:进入 nihao-ai.com 首页,点击右下角「在线客服」→「消息」→「开始对话」。
提交信息:直接发送您的用户名即可。用户名可在首页右上角头像 →「个人中心」中查看并复制。
发放时间:客服会在每天晚上 8 点前,通过在线客服聊天窗口发送兑换码。
创作平台升级

绘图视频创作站全面优化,高级创作功能更强、价格更低

最新图形生成站 qizuo.art 已完成体验优化:一个工作台连接图像、视频与叙事模型,从灵感、画面到成片都能快速完成,适合持续试错、批量产出和专业创作。

多模型创作:支持 GPT Image、Banana 4K、Grok Video、Veo Video,覆盖高清生图与视频生成。
高级功能:Prompt Optimizer 提示词优化、Director Studio 导演工作台、一键成片、镜头节奏控制和互动角色等功能全面优化。
低价体验:图片低至 ¥0.06/张,视频低至 ¥0.05/秒,注册即送积分,适合低成本高频创作。
立即体验:访问 qizuo.art 开始创作。
Codex API 优化

Codex API 配置流程升级,注册即送 1 美金体验额度

Codex API 站点 vibecodingboy 已完成新手流程和套餐配置优化,注册后即可获得 1 美金体验额度,多种包月套餐可选,快速上手更省心。

新手友好:优化注册、密钥和调用配置流程,跟着指引即可快速使用。
注册福利:注册即送 1 美金体验额度。
包月套餐:提供多种包月方案,按需选择。
市场低价:每 1 美元用量最低仅需 0.079 元,市场最低的源头价格,让您摆脱Codex 5小时限制,以更低成本体验官方 API 模型能力
站点地址:vibecodingboy.com
体验优化

新增谷歌登录,英文版适配同步优化

站点新增 Google 登录方式,注册和登录更方便;英文版页面适配同步优化,海外用户访问和使用更加顺畅。

Google 登录:支持使用谷歌账号快捷登录。
英文版适配:页面文案、布局和核心功能完成优化。

8月历史公告

以下为 8 月及更早的历史更新内容

8月福利

Claude 模型对所有 GPT 会员开放

现在所有 GPT 会员都可以在模型选择页面选择 Claude 模型,在 GPT 聊天页的模型切换处即可直接使用,无需额外开通 Claude 专属订阅。

使用位置:进入 GPT 聊天页,点击聊天框右侧显示“中/ 高 / GPT-5.6 Sol”等当前模型的位置,然后在弹出的模型列表中选择 Claude 模型即可。
开放范围:面向所有 GPT 付费会员开放。
新模型

4K 生图站加入 nano banana和grok 模型

4K 生图站现已加入 nano banana及grok 模型,用户可以根据创作需求在 GPT 模型与 nano 模型之间自由选择,兼顾不同风格与出图效果。

按需选择:同一生图站内可选择 GPT/grok 或 nano banana 模型。
适用场景:适合海报、头像、插画、商品图和高分辨率创作。
新增

主页增加 在线客服按钮

主页现已增加在线客服按钮,用户可以通过在线客服直接交流;智能客服会首先接待,并根据您的问题判断是否需要转人工处理。

入口更直接:在主页即可发起咨询,不需要额外寻找联系方式。
智能转接:常见问题由智能客服优先处理,复杂问题再转人工。
套餐

API 中转站推出 包月套餐

API 中转站现已推出包月套餐,额度更高更省,并支持通过一键配置程序快速配置 Codex / Claude Code App,减少手动填写参数的操作成本。

额度更高:包月套餐更适合高频调用、开发调试和长期使用。
一键配置:可使用一键配置程序接入 Codex / Claude Code App。

7月历史公告

以下为 7 月已发布的站点功能更新、模型上线、用量调整和使用提醒,保留在后方供用户回看。

福利

所有 Claude 订阅用户,用量提升 1.5 倍

即日起,本站所有 Claude 订阅账号的可用量整体上调至原来的 1.5 倍。无论你使用哪一档订阅,对话次数与高级功能的可用额度都同步提升,无需任何操作,已自动对全部用户生效。新购订阅套餐同样按上调后的更高额度计算。

自动生效:不需要重新领码或重新登录,当前账号即刻享受新额度。
计算方式:每 5 小时周期内的 Token 总用量,提升至原套餐的 1.5 倍。
新套餐生效:新购套餐已自动按升量后的额度启用,无需额外操作。
新功能

4K 生图创作站正式上线,高清出图一步到位

全新 4K 生图创作站现已开放,支持文生图与图生图,最高可直出 4K 超清分辨率。内置多种风格模型与画面比例预设,从概念草图到成品海报均可一站完成。在左侧模型选择器选择使用

4K 直出:原生支持最高 4096 分辨率输出,细节清晰,无需二次放大。
多任务并发:支持多任务同时生成 一次生成多图
支持参考图:可上传参考图 支持图生图
新模型

GPT-5.6 Sol 已正式上线

OpenAI 最新旗舰模型 GPT-5.6 Sol 现已更新 。在对话窗口右侧模型切换角标处即可一键切换使用。GPT-5.6 在编码、复杂推理、科学研究、网络安全和设计等领域带来显著性能提升,是当前最强 GPT 系列模型之一。

一键切换:对话窗口右侧模型切换角标处直接选择 GPT-5.6 Sol 即可使用。
能力亮点:编码基准领先、复杂任务处理更强。
立即可用:无需等待,已全量上线,所有订阅用户均可使用。
新模型

Grok 4.5 已正式上线

xAI 最新模型 Grok 4.5 现已入驻本站。在对话窗口右侧模型切换角标处即可一键切换使用。Grok 4.5 专为编码、Agentic 任务和知识工作深度优化,在真实工程场景中表现突出,速度快且性价比高。

默认调用:对话窗口提问默认调用最新 Grok 4.5
限时优惠

四合一会员限时大促,一份订阅畅享四大AI

即日起,本站推出四合一超级会员限时特惠:一份订阅同时解锁 ChatGPT、Claude、Grok、Gemini 四大顶级AI订阅版服务的尊享体验。无需分别注册、无需多份付费,单独订阅四家官方每月需近千元,现在开通四合一会员,仅需官方价格的一个零头,即可全部畅享。

一号通用:一个会员账号,四大AI全部解锁,随时切换、自由使用。
官方体验:号池内均为官方订阅账号,最新模型与高级功能同步可用。
限时特惠:活动期间开通立享超低折扣,价格随时恢复,先到先得。
开通方式:点击主页左上方升级会员 筛选器中选择四合一会员即可
新模型

Claude Fable 神话(Mythos)级模型,官方限时开放

Anthropic 推出了全新的 Claude 5 系列首款模型 Claude Fable 5,定位在 Opus 之上的全新「Mythos 神话级」梯队,是目前公开可用的最强 Claude 模型,在复杂推理、长任务规划和写作质量上均有显著提升。本站已第一时间接入,当前处于官方限时开放阶段,建议尽早体验。

如何使用:登录后在模型选择器中选择 Claude Fable 5 即可。
用量计算:Fable 的官方额度消耗非常高,按 10 倍率同步计费,使用 Fable 会更快达到个人用量上限(限额每 5 小时重置)。建议把它留给真正需要顶级能力的任务。
注意:本次为官方限时开放,若官方后续调整或取消开放,本站将同步官方说明并及时公告。
新模型

Claude Sonnet 5 全新模型已上线

7月1日,Anthropic 最新的 Sonnet 5 正式发布,面向编码、智能体和复杂任务全面升级,在保持 Sonnet 系列响应速度的同时大幅提升了代码与工具调用能力。现已在本站全量开放,作为日常主力模型使用性价比最高。

推荐用法:日常对话、写代码、处理文档默认选它。
额度友好:Sonnet 档位的用量消耗低于 Opus / Fable,同样额度能用更久。
新功能

Codex / Claude Code 一键接入

重磅功能:Codex 与 Claude Code 两大客户端现已支持一键接入。我们为你省去了全部复杂配置——下载客户端后,跟随引导走完几步,就能在自己的电脑直接用上 Codex / Claude Code。

接入方式:下载客户端 → 按引导逐步操作 → 完成,全程无需手动填写复杂参数。
调整

关于 GPT thinking 模型的用量调整

官方 GPT Plus 订阅(20 美金/月独享)的 thinking 功能额度为 200 次/周。为保障共享账号下所有用户的体验稳定,本月起将个人 thinking 用量设定为 50 次/天(合 350 次/周)。总可用量依旧高于自购一份 20 美金的独享订阅。

影响范围:仅 thinking / 深度推理类模型;日常 GPT 对话模型完全不受影响。
建议:简单任务交给普通模型,把每天 50 次 thinking 额度留给真正需要深度推理的问题。
通知

GPT 官网界面调整,模型选择器位置变更

GPT 官网近期更新了界面布局:模型选择器不再显示于页面左上角,而是移至对话输入框的右侧。功能与可选模型均无变化,仅为位置调整,找不到模型切换入口的用户请留意新位置。

新位置:模型切换入口现位于对话输入框右侧,点击即可展开模型列表。
功能不变:可用模型与切换方式均与之前一致,仅界面位置发生变化。
提醒

换车后继续对话的两种方式

更换账号(换车)后,请务必在新的对话框中提问,不要点击左侧历史记录——点击历史记录会跳转回该记录所在的原车号。如果想换号后接着之前的话题继续聊,用下面两种方式任选其一即可。

方式一 · 换号继续聊:点击齿轮按钮,选择「换号继续聊」,系统会自动切换账号并注入之前的上下文记忆。切换完成后可以先问一句"我们刚才聊了什么",确认记忆已带上。
方式二 · 下载对话续接:点击齿轮按钮,选择「下载当前对话」,再把下载的对话文件上传到新车号的对话框,让 GPT 根据前文内容继续聊即可。
更新

车队高级功能可用余量直观显示

车队页面新增高级功能余量面板:每个车队当前的高级功能(深度研究、文件处理、图像等)剩余可用量,现在以三个小圆点的形式直观展示,一眼即可判断余量充足与否,不用再靠试错摸索。

在哪里看:进入车队列表,每个车队卡片上即可看到各项高级功能的实时余量。
选车更聪明:上车前先看余量,优先选择目标功能余量充足的车队,体验更顺畅。

2024—2026 · 服务持续升级

GPT服务优化|对话隔离|防降智网关升级|
Gemini / Grok号池|Windows App

查看更多

AI Service Updates|Latest announcements
SITE UPDATES · OCTOBER 2026

October Site Updates

This month brings automatic GPT-6 updates, Work coding accounts, a GPT Go image-generation section, and a GPT image-quota notice.

All updates live Last updated: October 12 Direct connection to official sites

October Updates

Automatic GPT-6 updates, Work coding accounts, GPT Go image generation, and quota notes

Model update

GPT-6 is synced, with the latest model selected automatically

你好AI now includes the latest official GPT-6 model. No manual selection is needed in the GPT pool—the system automatically uses the latest available GPT-6 model.

Automatic selection: The system automatically uses the latest available GPT-6 model.
Ready to use: Open the GPT pool and start a conversation as usual.
New account type

Work (Codex) accounts added, optimized for coding and long tasks

The GPT pool now includes accounts labeled “Work”. These are the original Codex accounts and use Codex quota. They are optimized for programming, code analysis, and long-running tasks.

Open to all members: Work accounts are available to every GPT pool member. No separate Work plan is required.
Best for: Programming, code analysis, project development, and complex tasks that require sustained work.
Image generation

New GPT Go image-generation section, with dedicated capacity for image users

To support image-generation users, the pool now includes a dedicated GPT Go section. Scroll to the lower part of the pool to find and use it.

How to use: Open the GPT pool, scroll down, and select the GPT Go image-generation section.
Recommended: Use this section first when generating images for a more stable experience.
Usage notice

GPT image quota display issue, please rely on actual usage

The official GPT image-quota interface is currently returning incorrect data and cannot reliably show each account’s actual remaining quota. Because 你好AI connects directly to the official service, this issue also affects the quota information shown on our page.

Display note: The image quota shown as the three dots on an account card cannot currently be used as an accurate quota indicator.
What to do if you hit a limit: A message saying the current account has reached its image-generation limit does not mean your membership is unusable. Return to the homepage, choose another account, and start a new conversation before sending your prompt. Do not reopen the old conversation from history. After switching accounts and starting a new conversation, you can continue trying.
Recommendation: Image-generation users should use the GPT Go section below. We will keep improving account scheduling and do our best to provide a smooth experience even when accounts must be selected without quota information.

September Archive

Published updates from September, kept here for reference.

Latest model

GPT’s latest image model, Image 2.5, is now available

GPT Image 2.5 is now available in the GPT pool. Send an instruction to generate an image and the system will use Image 2.5 by default.

Ready to use: Open a GPT conversation and send an image-generation instruction—no extra model switch is needed.
Default model: Image-generation requests automatically use GPT Image 2.5.
New model

GPT-6 Model is now in the GPT pool

The GPT-6 model everyone has been waiting for is now available in the GPT pool. GPT-6 has just launched, and the official site has not yet rolled it out to the GPT Plus web app. To let you try it sooner, we connected GPT-6 through the official API.

How to use: On the GPT page, click the reasoning-strength selector (Instant / Medium / High) to the right of the chat box, then choose a GPT-6 series model from the dropdown.
Please note: Our pools have always connected directly to real official GPT Plus subscription accounts; regular models such as GPT-5.6 Sol are provided this way. GPT-6 is not yet available in the official GPT Plus web app, so it is connected through the official API for early access. Because the two connection methods differ, do not mix them in the same conversation or some messages may be missing. Start a new conversation when using GPT-6.
Member benefit

Monthly and higher plans receive 1,000 creative-studio credits

Everyone with any active AI monthly plan or higher receives a redemption code for 1,000 image and video creation credits.

How to claim: On the main site, click “Online Support” in the lower-right corner → “Messages” → “Start a conversation”.
Submit your details: Send your username directly. Find and copy it from the avatar in the upper-right corner → “Profile”.
Delivery time: Support sends the redemption code in the chat window before 8:00 PM each day.
Creation platform upgrade

Image and video studio fully optimized, with stronger tools and lower prices

qizuo.art has been upgraded for a smoother workflow: one studio connects image, video, and storytelling models, so you can move from an idea to finished content quickly. It is built for iteration, batch production, and professional creation.

Multiple models: GPT Image, Banana 4K, Grok Video, and Veo Video for high-resolution images and video generation.
Advanced tools: Prompt Optimizer, Director Studio, one-click video creation, shot pacing, and interactive characters.
Low-cost creation: Images from ¥0.06 each and video from ¥0.05 per second. Sign up to receive credits.
Try it now: Visit qizuo.art to start creating.
Codex API upgrade

Codex API setup is faster, with $1 in credits for new sign-ups

The vibecodingboy Codex API site now has a simpler onboarding flow and improved plan setup. New users receive $1 in credits after signing up, with monthly plans available for different workloads.

Beginner-friendly: Registration, key creation, and client setup are streamlined with step-by-step guidance.
Sign-up bonus: Receive $1 in credits when you register.
Monthly plans: Choose a monthly plan that fits your usage.
Competitive pricing: 1 USD of usage starts from just ¥0.079, giving you a lower-cost way to access official Codex API models.
Experience upgrade

Google Sign-In added, with improved English support

Google Sign-In is now available for faster registration and login. The English version has also been refined across layouts, copy, and core features for a smoother international experience.

Google Sign-In: Use your Google account to sign in quickly.
English version: Copy, layout, and core functions have been optimized.

August Archive

Historical updates from August and earlier

August benefit

Claude models are available to all GPT members

All GPT members can now select Claude models from the model picker on the GPT chat page. No separate Claude subscription is required.

Where to find it: In a GPT chat, click the current model selector to the right of the composer (for example “Medium / High / GPT-5.6 Sol”), then choose a Claude model.
Who can use it: Available to all paid GPT members.
New models

The 4K image studio now includes Nano Banana and Grok

Nano Banana and Grok models are now available in the 4K image studio. Choose between GPT and Nano models to match your creative needs and preferred visual style.

Choose as needed: Select GPT, Grok, or Nano Banana in the same image studio.
Great for: Posters, avatars, illustrations, product images, and high-resolution work.
New

An online support button is now on the homepage

You can now contact support directly from the homepage. The smart assistant handles common questions first and routes complex issues to a human when needed.

Direct access: Start a conversation from the homepage without searching for contact details.
Smart routing: Common questions are handled automatically, with human support for complex issues.
Plans

The API relay now offers monthly plans

Monthly API plans are now available with more quota and better value. One-click setup for Codex and Claude Code App reduces the need to enter parameters manually.

More quota: Monthly plans suit frequent calls, development, and long-term use.
One-click setup: Connect Codex or Claude Code App with the one-click configuration tool.

July Archive

Feature updates, model launches, quota changes, and usage notes published in July.

Benefit

All Claude subscribers get a 1.5× usage boost

Usage for every Claude subscription has been increased to 1.5× the previous amount. Conversation limits and advanced features are upgraded automatically for every tier, including new purchases.

Automatic: No new code or login is required; the higher quota is active immediately.
How it is measured: Total token usage in each five-hour window is 1.5× the original plan quota.
New plans: New purchases automatically use the increased quota.
New feature

The 4K image creation studio is live, with high-resolution output in one step

The new 4K image studio supports text-to-image and image-to-image generation, with output up to 4096 pixels. Style models and aspect-ratio presets take you from concept sketches to finished posters in one place.

Native 4K output: Up to 4096-pixel images with crisp detail and no second upscaling step.
Concurrent tasks: Generate multiple images and run multiple tasks at once.
Reference images: Upload a reference image for image-to-image generation.
New model

GPT-5.6 Sol is officially live

OpenAI’s latest flagship model, GPT-5.6 Sol, is now available. Switch models from the selector on the right side of the conversation window. It brings major gains in coding, complex reasoning, scientific research, cybersecurity, and design.

One-click switch: Select GPT-5.6 Sol from the model selector on the right side of the conversation window.
Highlights: Leading coding performance and stronger handling of complex tasks.
Available now: Fully rolled out to all subscribers.
New model

Grok 4.5 is officially live

xAI’s Grok 4.5 is now available. Switch from the model selector on the right side of the conversation window. It is optimized for coding, agentic tasks, and knowledge work, with strong real-world performance, speed, and value.

Default model: New conversations use the latest Grok 4.5 by default.
Limited offer

One subscription unlocks four leading AI services

For a limited time, the four-in-one super membership combines ChatGPT, Claude, Grok, and Gemini subscription features in one account. Skip four separate registrations and payments and enjoy all four services for a fraction of the official combined price.

One account: All four AI services are unlocked in one membership.
Official experience: Pool accounts are official subscriptions with current models and advanced features.
Limited-time price: Open during the promotion to lock in the discount before prices return.
How to activate: Click “Upgrade Membership” in the upper-left menu and choose the four-in-one plan.
New model

Claude Fable Mythos-level model is available for a limited time

Anthropic’s Claude Fable 5 is the first model in the Claude 5 series, positioned above Opus in a new Mythos tier. It delivers stronger complex reasoning, long-task planning, and writing quality. We connected it as soon as it opened for limited official access.

How to use: Select Claude Fable 5 in the model picker after signing in.
Usage multiplier: Fable consumes quota at a 10× rate, so it reaches the five-hour personal limit faster. Save it for tasks that need top-tier capability.
Availability: This is a limited official release and may change if Anthropic adjusts access.
New model

Claude Sonnet 5 is now available

Anthropic’s Sonnet 5 launched on July 1 with upgrades for coding, agents, and complex tasks. It keeps the speed of the Sonnet line while improving code generation and tool use, making it an excellent everyday choice.

Recommended use: Use it by default for everyday chat, coding, and documents.
Quota-friendly: Sonnet uses less quota than Opus or Fable, so the same plan lasts longer.
New feature

Codex / Claude Code one-click connection

Codex and Claude Code clients now support one-click connection. Follow the guided setup after downloading the client and use both tools on your computer without entering complex parameters manually.

Connection flow: Download the client → follow the guide → finish setup.
Adjustment

Updated usage for the GPT thinking model: new daily allowance

The official GPT Plus subscription includes 200 thinking uses per week. To keep shared accounts stable, the allowance here is set to 50 uses per day (350 per week), which remains higher than a standalone $20 subscription.

What is affected: Only thinking and deep-reasoning models; everyday GPT chat models are unchanged.
Tip: Use regular models for simple tasks and save the 50 daily thinking uses for problems that need deep reasoning.
Notice

The GPT website layout changed: model selector moved

The official GPT interface moved the model selector from the upper-left corner to the right side of the message composer. Available models and switching behavior are unchanged.

New location: The selector is now to the right of the message composer.
Function unchanged: Models and switching work as before.
Reminder

Two ways to continue a conversation after changing accounts

After switching accounts, always ask your next question in a new conversation instead of opening a history item, which returns to the original account. To continue the same topic, use either method below.

Method 1 · Switch and continue: Click the gear button and choose “Switch and continue”. The system carries over the previous context. Ask “What were we just discussing?” to confirm.
Method 2 · Download and resume: Choose “Download current conversation”, then upload the file in a new account’s chat and ask GPT to continue from the context.
Update

Fleet advanced-feature quota is now visible

The fleet page now shows remaining advanced-feature quotas—such as deep research, file processing, and image tools—as three small dots on each fleet card, so you can check availability at a glance.

Where to see it: Open the fleet list and check each fleet card for live remaining quotas.
Choose smarter: Check quota before boarding and pick a fleet with enough capacity for your task.

2024–2026 · Service keeps improving

GPT optimization · Conversation isolation · Anti-degradation gateway ·
Gemini / Grok pools · Windows App

View more