ChatGPT just rolled out a wave of changes, and these are not tiny cosmetic tweaks. The new ChatGPT 5.6 updates bring a redesigned interface, smarter settings, stronger voice options, new app permissions, and a major shift in how models are being packaged and priced. On top of that, ChatGPT Enterprise is turning into a real automation platform with skills, agents, company knowledge, scheduled tasks, and deeper workspace controls.
If you use ChatGPT regularly for writing, research, coding, operations, or team workflows, this matters. A lot of people will miss the most important changes because they are buried inside menus, permissions, and workspace settings. That is exactly where the real power is now living.
Here is what actually changed, what you should turn on immediately, and why GPT 5.6 could become a big competitive advantage once it is broadly available.
The new ChatGPT interface is much more action-oriented
One of the first changes you will notice is the updated UI. The plus menu has been redesigned, and that matters more than it sounds. It pushes ChatGPT away from being a simple chat box and toward being a control panel for multiple AI actions.
From that menu, you can now jump directly into things like:
- Image creation
- Web search
- Deep research
- OpenAI platform access
- Apps, plugins, files, and skills
This is a big usability shift. Instead of digging around for features, more of the core tools are available right where you compose your requests. If you keep files in your library, want to trigger a plugin, or need to reuse a custom skill, you can access that far more directly now.
You may also see MCP-related tools and other connected capabilities in this area depending on your setup. The direction is obvious: ChatGPT is becoming a central operating layer for your work, not just a place to ask questions.
The settings matter more than ever
This update is not just about new models. Some of the most useful improvements are hidden in settings, and if you skip them, you will not be getting the best version of the product.
Turn on dictation
There is now a setting to enable dictation inside the chat composer. If you like speaking ideas out loud, brainstorming quickly, or capturing tasks without typing, this is worth turning on immediately.
Dictation sounds minor, but it reduces friction. The easier it is to get information into ChatGPT, the more likely you are to use it for fast thinking, rough drafts, and workflow support.
Enable higher intelligence
One of the smartest new settings is higher intelligence. When enabled, ChatGPT can automatically step up its reasoning for more complex prompts.
That means you do not always have to manually choose a heavier model setting every time you ask something difficult. For everyday work, this can save time while still giving you stronger performance when complexity spikes.
If your tasks include:
- Deep analysis
- Complex coding
- Scientific or technical research
- Multi-step business reasoning
then this setting is especially useful.
Review app permissions carefully
Under Apps, ChatGPT now provides more explicit permission controls. You can decide whether it should always ask before taking certain actions, whether it can perform read actions, and whether it can take low-risk actions.
That includes things like:
- Reading from connected apps
- Making changes inside apps
- Sending or drafting emails
- Accessing information stored in external services
Honestly, this is a good move. Giving AI broad app access without checks is risky. OpenAI appears to be taking a more cautious route here, which makes sense if these tools are going to be used in real business environments.
Connector search should be on
In personalization and advanced settings, connector search is one of the most important options to enable. This helps ChatGPT search across connected tools and sources more effectively.
If you are using integrations across work apps, this can make a huge difference in how much context ChatGPT can bring into a task.
There are many more controls now
OpenAI has also added and expanded controls around:
- Data controls
- Storage
- Safety
- Security and login
- Parental controls
- Trusted contacts for account recovery
- Keyboard and input customisation
If you want to work faster and more safely, spend time in these menus. There are a lot of useful adjustments buried in there.
Voice mode quietly got better
Voice settings have improved too. ChatGPT can now auto-detect the language you are speaking, and more languages have been added. You can also choose which voice model to use.
If voice is part of your workflow, make sure you are using the advanced option. Otherwise, you may be stuck with an inferior voice experience without realising it.
For people who use ChatGPT while walking, commuting, outlining ideas, or handling quick task capture, this is one of those changes that makes daily usage noticeably better.
Where GPT 5.6 actually is right now
Here is where things get a little confusing. Inside the normal model selector, you may not see GPT 5.6 listed the way you expect. Instead, you will likely see the familiar tiering under GPT 5.5, including options like instant, medium, high, extra high, and pro.
That does not mean GPT 5.6 is not real. It means the rollout is being handled differently.
At the moment, GPT 5.6 is being previewed through the API and Codex for a limited group of trusted partners and organisations. Broader availability is expected, but it is not fully surfaced in the standard ChatGPT interface for everyone yet.
So yes, the update is here. It is just not fully exposed everywhere.
The three new GPT 5.6 models: Sol, Terra, and Luna
OpenAI is introducing three GPT 5.6 model variants, each aimed at a different balance of power, speed, and cost.
Sol
Sol is the flagship model. This is the one designed for the highest-end performance.
Terra
Terra is the balanced model for everyday work. The interesting part is that it is described as competitive with GPT 5.5 while being half the price.
Luna
Luna is the fast and affordable option. It is built to deliver strong capability at the lowest cost tier.
That model lineup is smart. Instead of forcing everyone into one expensive choice, OpenAI is clearly segmenting by use case:
- Sol for premium, high-stakes reasoning
- Terra for strong general-purpose performance
- Luna for speed and scale at lower cost
Why GPT 5.6 could be a major competitive advantage
Two things stand out with GPT 5.6: performance and economics.
OpenAI is positioning these models as outperforming strong alternatives on benchmark comparisons, including terminal-style evaluations and output, latency, and cost comparisons. There is also a strong emphasis on better cyber capabilities and stronger safeguards.
That last point matters. Powerful models are useful, but if they introduce security concerns or weak controls, adoption gets messy fast. OpenAI seems to be pushing the message that these new models are not just stronger, but also safer in how they are being released and governed.
The other major advantage is prompt caching. GPT 5.6 introduces more predictable prompt caching with:
- Explicit cache breakdowns
- A 30-minute minimum cache life
- Lower cost for repeated prompt structures
That is a big deal for developers, teams, and anyone building repeatable AI workflows. If you are reusing structured prompts at scale, caching can dramatically improve economics.
GPT 5.6 pricing
Pricing is listed per 1 million tokens across the three model sizes:
- Sol: $5 input, $30 output
- Terra: $2.50 input, $15 output
- Luna: $1 input, $6 output
That pricing structure makes Terra especially interesting. If it really delivers GPT 5.5-level competitiveness at half the cost, it could become the default sweet spot for a huge number of use cases.
For API users, this is worth tracking closely through the OpenAI platform. If you are already building workflows or tools, these economics may change what model you deploy by default.
ChatGPT Enterprise is becoming a serious work platform
If you are using ChatGPT for actual business operations and not just occasional prompting, ChatGPT Enterprise looks increasingly compelling.
The difference is not subtle. Enterprise unlocks a much broader layer of control, automation, and shared workspace capability.
Reusable skills
Inside your workspace, you can now create skills. These are reusable behaviours or output patterns that can be built in a few different ways:
- Create with chat
- Create with the editor
- Upload from your computer
This is huge for consistency. If you ever get ChatGPT to produce an output you love, you can turn that into a repeatable skill instead of rebuilding the same prompt over and over.
For example, if you have a formatting workflow, a tagging system, a reporting style, or a recurring analysis pattern, you can save that logic and reuse it.
That shifts ChatGPT from ad hoc prompting to systematised execution.
24/7 AI agents
Enterprise also includes AI agents that can run continuously. There are templates in the team directory for common business roles and workflows, including things like:
- Chief of staff
- Customer reply drafter
- Product planning support
- Marketing strategy support
- Knowledge research
You can use templates to get started quickly, or create your own agent for very specific jobs like sending daily briefs, answering team questions, or handling repeated internal processes.
This is where the platform starts looking less like a chatbot and more like an operating system for AI work.
Workspace-specific apps
Enterprise workspaces also support app configurations that go beyond what is available in regular ChatGPT. If your goal is real automation, this matters a lot.
More apps means more reachable context, more actions, and more end-to-end workflows without manually hopping between systems.
Company knowledge across tools
One of the strongest Enterprise capabilities is the ability to create a shared company knowledge layer by connecting business tools.
This can include services like:
- Slack
- Gmail
- Google Drive
- Outlook
- Calendar
Once connected, ChatGPT can operate with richer company context. That makes a massive difference for analysis, recommendations, drafting, internal support, and decision assistance.
If you are asking ChatGPT to help with business decisions but you are not giving it business context, you are leaving an enormous amount of value on the table.
Scheduled automations are one of the most practical upgrades
Another standout feature is scheduled tasks. ChatGPT can now handle recurring automations like reminders, monitoring, and updates.
Built-in presets include things like:
- Daily briefings
- Email monitoring
- Sports recaps
- Sales monitors
- Concert alerts
- Awards tracking
- Weekend ideas
And if that is not enough, you can simply describe what you want it to do.
There is also support for folding apps into automations by referencing them directly. That means workflows can become much more actionable instead of staying stuck at the reminder level.
This is one of the clearest signs that ChatGPT is moving toward active assistance rather than passive response.
Admin controls and hidden workspace settings
For teams, Enterprise includes deeper controls around plugins, permissions, roles, and workspace-level settings. This gives administrators a way to standardise how ChatGPT is used across an organisation.
That includes the ability to manage:
- Available plugins
- User permissions
- Role-based settings
- Agent configuration rules
- Operational customisations
There are also many niche settings most people would never discover unless they explored the workspace controls closely. Some are tiny, but together they can meaningfully change how smoothly your systems run.
If you manage a team, this matters because you no longer have to trust every person to configure everything correctly on their own. You can shape the environment centrally.
What to do right now
If you want to get the most out of these ChatGPT updates, here is the practical checklist:
- Open your settings and enable dictation.
- Turn on higher intelligence.
- Review app permissions so actions are gated the way you want.
- Enable connector search in advanced settings.
- Check voice settings and switch to advanced voice if needed.
- Explore the new plus menu for direct access to apps, files, and skills.
- If you use ChatGPT for business, seriously consider Enterprise.
- Inside Enterprise, test skills, agents, and scheduled automations.
- Connect business tools so ChatGPT has real organisational context.
- Keep an eye on GPT 5.6 API rollout and pricing if you build with OpenAI.
Related resources
For official product and developer information, check the OpenAI Platform.
If your site has related content, this article would pair well with internal links to pages about hidden ChatGPT settings, prompt engineering, AI automations, or ChatGPT for teams.
FAQ
Is GPT 5.6 available inside regular ChatGPT yet?
Not fully for everyone. GPT 5.6 is currently being previewed through the API and Codex for select trusted partners and organisations, with broader availability expected later.
What are the new GPT 5.6 models?
The three GPT 5.6 models are Sol, Terra, and Luna. Sol is the flagship model, Terra is the balanced everyday model, and Luna is the fast, affordable option.
Which new ChatGPT settings should I turn on first?
Start with dictation, higher intelligence, connector search, and advanced voice. Then review app permissions so ChatGPT behaves the way you want when interacting with connected tools.
What is the benefit of higher intelligence mode?
Higher intelligence allows ChatGPT to automatically use stronger reasoning for more complex prompts. It helps you get better responses without manually changing settings every time.
Is ChatGPT Enterprise worth it?
If you use ChatGPT for real work, automation, or team collaboration, Enterprise appears significantly more powerful than the standard experience. Skills, agents, scheduled tasks, company knowledge, and admin controls make it a very different product.
What are ChatGPT skills?
Skills are reusable workflows or behaviours you can create from prompts, editors, or uploaded files. They let you save a useful output pattern and apply it again without rebuilding it from scratch.
What can ChatGPT scheduled tasks do?
Scheduled tasks can handle reminders, daily briefs, monitoring, and update-based automations. They also support preset workflows and can incorporate connected apps into the automation.
The bottom line
These ChatGPT 5.6 updates are not just about one new model number. They represent a broader shift in what ChatGPT is becoming.
The consumer side is getting more flexible with better settings, stronger voice, and easier access to tools. The model layer is getting more segmented and cost-aware with Sol, Terra, and Luna. And the Enterprise side is evolving fast into a serious AI workspace for automation, team operations, and connected business context.
If you have been using ChatGPT like a simple chatbot, this is a good moment to rethink that. The people who get the most out of this next phase will be the ones who treat it like infrastructure.
If you found this helpful, share it with someone who uses ChatGPT daily, leave a comment with the feature you are most excited about, and explore more AI workflow guides on the site.
