ChatGPT Triples Custom Instructions Limit: What It Means for You
OpenAI expanded ChatGPT's custom instructions field from 1,500 to over 5,000 characters, allowing much richer persistent context in every chat.
What changed in ChatGPT’s custom instructions?
OpenAI quietly expanded the character limit on ChatGPT’s custom instructions field, more than tripling it from 1,500 characters to over 5,000. The setting lives under Personalization in ChatGPT’s settings menu, and it lets users define background context, preferences, and rules that get applied to every new conversation automatically. The change means people can now paste in far more detailed personal or professional context without hitting a wall partway through.
TL;DR
- OpenAI raised the custom instructions limit from roughly 1,500 characters to more than 5,000, giving users room for significantly more persistent context.
- The update lives in ChatGPT’s Personalization settings, the same place users have always gone to tell the model how to behave or what to know about them.
- Longer context documents now fit directly, so users who previously had to trim or summarize personal background files can paste much more of it in as is.
- Every message benefits from this context automatically, since custom instructions apply to all new chats without needing to re-explain who you are each time.
- Running more context through every message has a cost, which is likely why OpenAI kept the limit tight for as long as it did before expanding it now.
- The update is easy to miss since it’s a settings tweak rather than a flashy new feature, but it changes how much useful personalization is practical.
How does the custom instructions feature actually work?
Custom instructions are a standing block of text that ChatGPT references before responding to anything you type. Instead of re-explaining your job, your writing style, or your project details in every new conversation, you write it once in settings and the model treats it as background context for every future chat.
Typically the field is split into prompts like “What would you like ChatGPT to know about you?” and “How would you like ChatGPT to respond?” People use it for things like professional role, tone preferences, recurring projects, or personal facts that make answers more relevant without extra prompting.
The expanded limit does not change how the feature works mechanically. It changes how much you can put into it. A 1,500-character cap was tight enough that anyone with a genuinely detailed personal profile, resume-style background, or set of working preferences had to cut it down to fit. At over 5,000 characters, there is room for a much fuller picture: history, goals, communication style, tools you use, and constraints you want respected.
Why did the limit stay so small for so long?
Every custom instruction gets sent to the model as part of the input for every single message, whether it is relevant to that specific question or not. More characters in that field means more tokens processed on every request, which adds cost and latency at scale across millions of users. Keeping the field capped at 1,500 characters was a straightforward way to control that overhead while still letting people set basic preferences.
The jump to 5,000-plus characters suggests OpenAI has decided the tradeoff is now worth it, likely because model costs have dropped and the personalization payoff is clear. It is a small technical change with an outsized practical effect for anyone trying to make ChatGPT feel like it actually knows them.
What can you fit in 5,000 characters that you couldn’t before?
The realistic use case here is what some AI practitioners call a personal context file, sometimes referred to informally as a “life” or background document: a written summary of who you are, what you do, your working style, your goals, and recurring context you want an AI assistant to have on hand. Before this update, anyone with a document like that had to aggressively trim it to survive the 1,500-character cap, often cutting exactly the detail that made it useful.
With more than 5,000 characters available, a moderately detailed personal or professional profile can now go in largely intact. That might include:
- A summary of your role, industry, and current projects
- Preferred response style (formatting, tone, level of detail)
- Recurring tools, terminology, or shorthand specific to your work
- Standing constraints, like avoiding certain phrasing or always including specific caveats
For people who maintain this kind of context in an external document (a Google Doc or similar file used to store background info for reuse across AI tools) it may now be possible to paste that content directly into custom instructions rather than summarizing it down first.
Is this update worth acting on?
Seven tools to build an app. Or just Remy.
Editor, preview, AI agents, deploy — all in one tab. Nothing to install.
If you already use custom instructions and previously had to cut content to fit the old limit, yes. Go back into Personalization settings, check the current character count, and consider whether there is more useful context you can now add. Because the instructions apply automatically to every new chat, filling in more detail once tends to pay off repeatedly rather than requiring ongoing effort.
If you have never used custom instructions, this is also a reasonable moment to start. The barrier to writing a useful profile is lower now that there is meaningfully more room, and the feature requires no technical setup, just a settings field and some honest description of how you want the model to work with you.
The update is not dramatic on its own. It is a quiet backend change rather than a new capability. But for anyone treating ChatGPT as a daily working tool rather than a one-off question box, having more room for persistent context is a real, if unglamorous, improvement.
Frequently Asked Questions
Where do I find custom instructions in ChatGPT?
They are located under Settings, then Personalization. Look for the custom instructions option, which typically includes fields for background information and response preferences.
Does the new character limit apply to all ChatGPT plans?
The transcript and available reporting describe the change as a general update to the custom instructions field rather than a tier-specific feature, though users should check their own account settings to confirm current limits, since rollout timing can vary.
Will longer custom instructions slow down responses?
Adding more context does mean more text is processed with every message, which can have a marginal effect on response time or cost behind the scenes, but it does not require any extra action from the user and is handled automatically by ChatGPT.
What should I put in custom instructions if I’m starting from scratch?
Focus on stable, reusable facts: your role or profession, how you prefer answers formatted, recurring projects or terminology, and any consistent constraints on tone or content. Avoid highly time-sensitive details that will go stale quickly.
Is this the same as ChatGPT’s memory feature?
No. Custom instructions are a fixed block of text you write yourself and control directly. Memory is a separate system where ChatGPT can retain facts from conversations over time. They can work together, but custom instructions are static until you edit them, while memory updates based on ongoing chats.