GPT 5.6 is out TOMORROW!

GPT 5.6 Sol along with Terra and Luna will launch on Thursday. Happy days! They're expanding preview access globally right now. Opening Eyes next model is coming.

Hopefully it's on par or better than Fable 5. If not, no worries at all. But today I'm going to show you how to prepare. This is a GPT 5.6 prep machine.

So GPT 5.6 launches publicly this Thursday. It's the most capable model Open Eyes has ever shipped and it will supercharge any agent operating system you have in. And here's exactly how to be ready on day one when it comes out. So if you're not familiar with this, basically Open Eyes have announced GPT 5.6 with three models.

Sol the flagship, Terra the lower cost and Luna the fastest and cheapest. Now preview access is expanding globally. The public launch is Thursday. Most people start figuring out on launch day you're gonna do something different.

You're gonna prep the machine this week so that the moment GPT 5.6 drops it plugs straight into your automations and starts working whilst everyone else is still reading the announcement. And the prep is really simple. I'm gonna walk you through exactly how to get the most out of it and how to really prepare for this. And the thing that I would say here is that there's four moves to get the most out of GPT 5.6.

So first thing you want to do is write down everything that you do day-to-day. For a few days or just today log every task you touch because you can't automate work you can't see. Then from there time audit it. So put a number on each task, how long, how often and now you know exactly where your hours will go.

From there minimize the tokens. So before you point a pricier or powerful model at everything, cut the waste. Free models for the everyday stuff and also you can use something like Headroom, Ponytail or RTK or even Caveman and these are free open-source projects you can just give to GPT 5.6 or Codex right now and then you're ready the day it comes out. And then you're gonna build an agentic operating system around the audit, around the stuff that you time and figure out you need to automate.

So for example I actually have a to-do list or I used to have a to-do list and now I just plug it into this system so that each day I can see what I've done and I can go back to it and I can just tick off the task. Super useful and it just saves me clicking into another app or another tool, losing track or anything else. Now also every time that I have a workflow that I'm like right why am I wasting time with this, why am I going into my website doing keyword research or why am I going into Heygen to do the avatar videos when I can just have it inside one system. That's what this is about.

So you're gonna write down everything you do, log every task, time audit it so see how long and how often you do it. Minimize the tokens inside any GPT models you use because then you can use those tokens in the future and you can get more out of GPT 5.6. If you saw anything or if you learned anything from Fable 5 it is that you can rinse through the tokens and the credits very quickly indeed my friend. So for example we've been using GPT, sorry Fable 5 today and you can see here that we've had to start, we've gone through the weekly Fable allowance and that resets on Saturday and so we can't really use Fable 5 for free again until that limit.

Now if you use the token playbooks I talked about you can save a lot of tokens to squeeze more out of the model before you hit the limits. So what to do this week? Today start keeping a log. I would every day just track what you're doing at least for one week because every time you start a task, writing an email, researching a keyword, editing a video, whatever it is, answering a client, write one line what it was and roughly how long.

Don't judge it, just capture it because by day three you'll have a raw map of your actual week and then from there tag each line with times of frequency. So go down the list, mark minutes per task and how many times a week you do it, multiply it, the tasks at the top, big number, done often, are you gold? That's where an agent would pay for itself. From there sort the list into three buckets.

A machine could do this today. A machine could help but needs me. Only I can do this right? And I would say if you're really honest with yourself most people actually shocked how much lands in the machine could do this today.

I was very shocked when I started doing this. And then minimize the tokens before you scale. Now before you point any price here, model at the automations, cut the waste. You want to run for example caveman, ponytail RTK, those playbooks which you can get on GitHub.

So for example we type in caveman here, you can see that we can get access to GitHub with the playbook for minimizing our work and then you can just go into codex and just drop that in. Pretty good. Now if you're wondering okay which models are coming out here, so you got Sol which is a flagship, lower cost from Terra, Luna is the fastest and the cheapest. If you're wondering what the benchmarks are you can see them here.

Sol is the most powerful, most capable and also I would change the effort level depending on what you're doing. So if you're doing like something quick like just going back and forth in the chat, then you can use Luna. If you're doing something that is really required, you know like coding out a dashboard, then you would go with Sol. So depending on the task you want to switch the model so that you don't waste the tokens that you have.

Now there's also a bunch of things you can do to cut the waste when these models arrive so you don't rinse through your credits. And the reason I keep going on about this is because when I've used codex in the past you can run out of tokens like so quickly, even on the most expensive plan. So you want to use like free or local models for the everyday stuff, you know the the 90% of tasks that you don't really need a frontier model for. Then catch the free windows, so for example like Tencent's HY3 is free on the API for two weeks.

You could plug it into agents to do the heavy work. It's also available for free on Noose portal for two weeks. And then you want to diet the paid replies. So when you do call a paid model, the Cayman token diet cuts what it says by 60 to 75% on the output tokens.

That means the answer is shorter but you get less output tokens and therefore you can serve everything. We actually tested out with Fable 5 and we got 69% fewer output tokens which made it 37% cheaper to use. Pretty amazing. And then die all the effort.

So GPT 5.6 lets you choose reasoning effort per task, simple job, low effort, fewer tokens. Match effort to the work and the flagship stays more affordable. You might also say you know why why minimize now? Well honestly later is when the build is already trained you to use AI less or later is when for example you run out of tokens on the credit.

If you minimize now and you build the habit for future frontier models, the machine keeps every future model cheap not just today but in the future. So for example our agent operating system as you can see right here, when we're using like basic models we'll just use Omniroot which is a free way to code for free. We can preview everything we've built, we've got the workspace over here, we can code in for free inside here. That's what you want.

So for example as well codecs. Recently we added HY3 inside codecs so we can code for free with codecs using Tencent's HY3 which is available on the API. So when you start implementing stuff like that you can just save so many tokens and you can squeeze way more and get way more done way more efficiently. So this is pretty powerful stuff and some people say I'll get ready when GPT 5.6 actually launches but it's only 24 hours away probably less than that right now.

So you want to start thinking about this now. Other people say well a more powerful model means a bigger bill but that's only if you skip the token minimize techniques I talked about before. And other people say well the model is a thing. Get the best model and you win.

Actually the model is the engine. The system around it is what wins. So GPT 5.6 is great but you want great systems built around it because bear in mind like if you look at what happened with Fable 5 like GPT 5.6 could get pulled after like 24 hours after a week and then you just totally lost. So what you want to do is have a great system built around this that gets better no matter what model drops and that way latency of all the powerful systems that we've built we can get way more out of our agents without having to worry about model.

The system gets better even if the models don't. So that helps you prepare in case for example GPT 5.6 gets pulled just like it happened with Fable 5. Now if you want the machine that I've shown you today with everything built in then you can get that inside the AI Profitable Boarding. Link in the comments description or go to the AIProfitableBoarding.com.

We've got the video agent, the SEO agent, we've got the memory system, we have codecs with free APIs, we have for example OmniRoot which is a free AI coder, local models ready to go, everything built into one beautiful system. If you want to get that it's all available inside the AI Profitable Boarding. Link in the comments description or just go to the AIProfitableBoarding.com. Inside the community you can post questions, get help and support in real time and I personally answer them with video tutorials.

Inside the classroom you can get access to all of our best trainings on this stuff and you can go from complete beginner to expert using this course. Also inside the new daily updates every time something useful comes out like for example HY3 and OmniRoot are free we create a tutorial on it. Every time for example we find new ways to reduce tokens we create a tutorial on it and every day I spend about three to four hours improving the AgentOS system so as soon as GPT 5.6 comes out we'll be using it to improve the system and make it even better. And you see it's updated daily, you can get the full zip file to install and that's all available.

Plus we have four weekly coaching calls where you can get help and support in real time, share your screen, meet other members. Inside the map you can meet people in your local area who are building with AI agents like you and that's all available, link in the comments description or just go to the airprofitable.com. Thanks for watching.