NOTICE: By continued use of this site you understand and agree to the binding Terms of Service and Privacy Policy.
如果你想要看中文文档，请往下滚动网页。
If you want to view the Chinese document, please scroll down the webpage.
Due to some special reasons, this script will no longer be updated and will be transferred to talk-to-gpt-4-mobile for code updates, together with Unintendedz.
gpt-4 and gpt-4 mobile are exclusively available for plus users. If you don't have a plus account, don't download this script unless you're interested in the gpt-3.5 and gpt-3.5 mobile models.
You can use GPT-4 Mobile in the same way as you would use GPT-3.5.
"Enhance your conversations" with the gpt-4-mobile model using a userscript. You can achieve this by combining the resources available at https://github.com/Unintendedz/talk-to-gpt-4-mobile and https://greasyfork.org/zh-CN/scripts/464284-chatgpt-is-gpt4-by-default-plus-only-available. By utilizing the Tampermonkey script, you can engage in conversations with the gpt-4-mobile model, removing the restriction of 25 messages every 3 hours, and set GPT-4 Mobile as the default model. Furthermore, the gpt-3.5 mobile option has also been incorporated.
Currently, a permission check has been added. Users without a "plus" subscription can easily switch between 3.5 mobile and 3.5 versions. However, for "plus" users, I have disabled 3.5 mobile. If you really want to use mobile 3.5, you can directly enter https://chat.openai.com/?model=text-davinci-002-render-sha-mobile in the address bar and press enter. This URL will allow you to use 3.5 mobile for conversations.
Thank you to both authors for the majority of the code contributions.
My job in this script was simply to stitch together two scripts and make them work seamlessly together. However, if you need to use the two scripts separately, perhaps you can download each of them separately as well.
This project aims to optimize and improve the user experience of OpenAI's chat model, addressing two main issues:
Default model selection: This Tampermonkey script automatically sets the user's specified default chat model, eliminating the need to manually select a model every time a new conversation is started. Users can enhance their efficiency by setting a default model based on their own needs and preferences.
Inclusion of mobile models: This script enables users to utilize OpenAI's mobile models (such as GPT-4 Mobile v2) even on non-mobile devices. This expands the range of model options, allowing users to choose a more suitable model based on their network environment and computational capabilities.
This project is written in JavaScript and runs through the Tampermonkey extension, enabling convenient and efficient usage directly within the browser.
Tips: GPT-4 and GPT-4 Mobile are available exclusively for Plus users.
Tips: This tool also supports GPT-3.5 mobile.
Here are some observations and speculations found online:
Some users have noticed that GPT-4 Mobile V2 is used when creating chats in mobile applications, and they believe that this version is not as intelligent as the original GPT-4.
A few users have noticed that this model is also labeled as GPT-3, leading them to suspect that this version might just be another replica of GPT-3.5.
There are users who view this as deceptive behavior. They believe that when creating conversations on both mobile and web platforms and selecting GPT-4, there are noticeable differences. They perceive it as a way to consume GPT-4 credits while actually using the GPT-3 model.
Some users have observed that GPT-4 Mobile V2 appears to lack a limiter, speculating that this might be similar to the advancements between the older and newer versions of GPT-3.
Others speculate that OpenAI may have managed to shrink the model while maintaining its performance, making it less costly to operate.
However, these are all user speculations and observations, as OpenAI has not provided any official explanations.
Reference link: gpt4 vs gpt4 mobile v2
Here is the table summarized by GPT-4.
| Number | Model Name | Description | Speed (0-100) | Inference Ability (0-100) | Simplicity (0-100) | Drawbacks |
|---|---|---|---|---|---|---|
| 1 | GPT-3.5 Mobile | Fastest model, suitable for most daily tasks | 100 | 60 | 40 | Suitable for daily tasks, limited inference capability for complex tasks |
| 2 | GPT-3.5 | Fast speed, suitable for most daily tasks | 90 | 60 | 40 | Suitable for daily tasks, limited inference capability for complex tasks |
| 3 | GPT-4 Mobile | Optimized model, better for innovative and advanced reasoning tasks | 80 | 100 | 80 | Slower speed, but strong inference capability |
| 4 | GPT-4 | Most capable model, suitable for innovative and advanced reasoning tasks | 70 | 100 | 80 | Slower speed, but strong inference capability |
| 5 | GPT-4 Browsing | Experimental model that knows when and how to browse the internet | 60 | Unknown | Unknown | Experimental model, effectiveness unknown |
| 6 | GPT-4 Plugins | Experimental model that knows when and how to use plugins | 50 | Unknown | Unknown | Experimental model, effectiveness unknown |
Due to modifications in some registered script information, it is possible that customers who installed the script early may encounter difficulties in updating. If you encounter this problem, clearing the project's cookies will enable you to update or reinstall this script.
因为一些特殊原因，此脚本不再更新，将转到 talk-to-gpt-4-mobile与 Unintendedz 一起更新代码
gpt-4 和 gpt-4 mobile 仅限 plus 用户使用。 没有 plus 账号的别下这个脚本，除非你是冲着 gpt-3.5 和 gpt-3.5 mobile 模型来的。
在本文中的无限指的是你可以像使用 gpt-3.5 一样使用 gpt-4 mobile
通过使用用户脚本，您可以提升与 gpt-4-mobile 模型的对话体验。您可以将位于 https://github.com/Unintendedz/talk-to-gpt-4-mobile 和 https://greasyfork.org/zh-CN/scripts/464284-chatgpt-is-gpt4-by-default-plus-only-available 的资源结合起来实现这一目标。通过利用 Tampermonkey 脚本，您可以与 gpt-4-mobile 模型进行 "对话"，消除每3小时25条消息的限制，并将 GPT-4 Mobile 设置为默认模型。此外，还包含了 gpt-3.5 mobile 选项。
目前添加了判断权限，没有 plus 的用户可以方便的切换 3.5 mobile 和 3.5 ，不过对于 plus 用户来说，我将 3.5 mobile 屏蔽了，如果你确实想要使用 mobile 3.5 那么可以直接在地址栏中输入 https://chat.openai.com/?model=text-davinci-002-render-sha-mobile
此网址敲下回车即可使用 3.5 mobile 进行对话。
感谢两位作者为大部分代码贡献而付出的努力。
我在这个脚本中的工作是简单的缝合了两个脚本让它们融洽的在一起工作，但是如果你需要使用两个脚本分别控制，那么也许你可以分别去下载这两个脚本。
本项目旨在优化和改善OpenAI聊天模型的用户体验，主要解决了以下两个问题：
默认的模型选择问题：该油猴脚本能够自动设置用户指定的默认聊天模型，避免了每次启动新对话时都需要手动选择模型的麻烦。用户可以根据自己的需求和偏好设置默认模型，提高使用效率。
添加移动模型的问题：该脚本能够使用户在非移动设备上也能使用OpenAI发布的移动模型（如GPT-4 Mobile v2）。这增加了模型的选择范围，使用户可以根据自己的网络环境和计算能力选择更适合的模型。
此项目基于JavaScript编写，通过油猴（Tampermonkey）插件运行，可以直接在浏览器中使用，方便快捷。
tips: gpt-4 和 gpt-4 mobile 仅限 plus 用户使用。
tips: 本工具也支持式 gpt-3.5 mobile
以下是网上的一些观察和猜测：
有些用户发现在移动应用中创建聊天时会使用GPT-4 Mobile V2，他们认为这个版本并不像原版的GPT-4那样智能。
一些用户注意到这个模型也被标识为GPT-3，他们怀疑这个版本可能只是GPT-3.5的另一个复制品。
有用户认为这是一种欺诈行为，他们认为在手机上和在网页上新建对话时提出同样的问题，选择GPT-4，会看到明显的区别。他们认为这只是另一种消耗GPT-4信用的方式，实际上使用的是GPT-3模型。
也有用户观察到GPT-4 Mobile V2似乎没有限制器，他们猜测这可能与GPT-3的旧版和新版之间的进步类似。
另一些用户认为OpenAI可能已经设法使模型变小，同时保持模型的性能，这样运行起来就不会那么昂贵。
然而，这些都只是用户的猜测和观察，OpenAI尚未对此做出任何官方解释。
下面是 gpt-4 总结的表格
| 序号 | 模型名称 | 描述 | 速度(0-100分) | 推理能力(0-100分) | 简洁度(0-100分) | 缺点 |
|---|---|---|---|---|---|---|
| 1 | GPT-3.5 Mobile | 最快的模型，适用于大部分日常任务 | 100 | 60 | 40 | 适用于日常任务，但对于复杂任务的推理能力有限 |
| 2 | GPT-3.5 | 速度很快，适用于大部分日常任务 | 90 | 60 | 40 | 适用于日常任务，但对于复杂任务的推理能力有限 |
| 3 | GPT-4 Mobile | 优化后的模型，对于需要创新和高级推理的任务更优 | 80 | 100 | 80 | 速度较慢，但推理能力强 |
| 4 | GPT-4 | 最有能力的模型，适合需要创新和高级推理的任务 | 70 | 100 | 80 | 速度较慢，但推理能力强 |
| 5 | GPT-4 Browsing | 知道何时如何浏览互联网的实验性模型 | 60 | 不明 | 不明 | 实验性模型，效果未知 |
| 6 | GPT-4 Plugins | 知道何时如何使用插件的实验性模型 | 50 | 不明 | 不明 | 实验性模型，效果未知 |
因为修改一些脚本注册的信息，所以可能导致早期安装的客户无法更新，如果遇到此问题，清除本项目的cookie 即可更新或者重新安装此脚本。
微信号 onepisYa
下面是我的微信 视频号，感兴趣的可以扫码关注。
Here is my WeChat Video account. If you're interested, you can scan the code to follow me. If you use wechat.
Rating: 0