Kind: captions
Language: en
You're probably still trying to figure
out if GPT5 was even worth the upgrade.
I mean, OpenAI hyped it up. It finally
launched in 2025 and then it felt
colder, more robotic, less helpful than
GPT4 in some ways. Users were frustrated
and even Sam Alman admitted they totally
screwed up some things with that launch.
But here's where it gets interesting.
Altman's already moved on. He's talking
about GPT6. And according to him, this
is where the real revolution happens. So
in this video, I'm breaking down
everything we know about GPT6,
the features that could actually
transform how we use AI, when it's
coming, and how you'll get access to it,
especially if you're in the US. We're
talking long-term memory that remembers
you across sessions, AI that can
actually execute tasks autonomously, and
a release timeline that's way shorter
than you'd expect. First up, let's talk
about why GPT5 fell flat and how that's
setting the stage for what OpenAI is
planning next. From GPT5's stumbles to
GPT6's promise, GPT5 wasn't the game
changer everyone hoped for.
When it launched in August 2025, users
immediately noticed the responses felt
impersonal and colder than GPT4.
OpenAI scrambled, even offering GPT4
outputs again to keep people happy.
Altman admitted the launch was rocky,
but instead of dwelling on it, he
immediately shifted the conversation to
GPT6,
claiming this is where the real
revolution happens. And based on what
he's been saying publicly, they're not
just trying to make a bigger model.
They're trying to make a smarter, more
useful one that actually remembers you
and can execute tasks autonomously.
Sam Alman's vision for GPT6.
Sam Alman has been surprisingly
transparent about what he wants GPT6 to
be. His message is clear. Size isn't
everything. It's about usefulness. The
headline feature, memory.
People want memory, he said repeatedly.
And he's right. Every time you start a
new chat with chat GPT, it's like
talking to someone with amnesia. You
have to reexplain everything. It's
exhausting.
GPT6 aims to fix that.
Altman envisions an AI that remembers
you across sessions, your preferences,
your style, your ongoing projects, less
like a tool, more like a long-term
companion that understands your needs
over time. But here's where it gets
really interesting.
Alman hinted at letting users shape the
AI's personality and worldview.
Right now, ChatGpt tries to stay
neutral, but he gave this example. If
you want your AI to be super woke or
lean conservative, it should adapt.
Formal and professional or casual and
friendly, it should match your vibe.
This is a massive shift from
oneizefits-all to true personalization.
You get to customize it to match your
values and communication style.
It's February again and while most
people are making resolutions, the smart
ones are already mastering the one skill
that matters most in 2026. AI.
From a simple text model in 2019 to
detecting diseases and automating work
by 2025, AI has come insanely far. In
2026, this is when it peaks. Last chance
to get on board. So why not reclaim
those six years in just 2 days?
That's why I'm excited to tell you about
Outskll. They've trained over 10 million
people, and they're running their 2-day
AI mastermind this Saturday and Sunday,
10:00 a.m. to 7:00 p.m. EST, completely
free as part of their new year
upskilling fest. 16 hours where you
build AI agents, automate workflows,
connect tools like notion and sheets.
People from this are making 2 to 3K
weekly with AI services. You get the AI
prompt bible,
AI profit road mapap,
2026 AI survival hackbook,
and a personalized toolkit if you attend
both days. Seats are filling up fast, so
hit that link in the description to grab
yours and join their WhatsApp community.
You'll get all the session links and
updates there.
Key feature one, long-term memory and
personalization. Let me break down this
memory feature because it's honestly the
most exciting part of GPT6.
Up until now, even with GPT4 and 5, the
AI only remembers what's in your current
chat session, maybe a few thousand words
of context.
Start a new chat, it's like hitting
reset on the AI's brain. Everything you
discussed before is gone. GPT6 is
expected to completely change this.
OpenAI is working on giving the model
true long-term memory that persists
between sessions.
Imagine this. GPT6 recalls that you're a
Python developer who prefers concise
code examples. It remembers you asked
for cooking tips last week. It knows
about that trip you're planning. Instead
of repeating yourself constantly, the AI
picks up right where you left off, even
if it's been days or weeks since your
last conversation. Think about what this
means practically.
No more repetitive reintroductions. If
you're working on a novel, GPT6 could
remember your characters, plot points,
and writing style from previous
sessions.
You won't need to upload the same
documents or explain the same context
every single time, but it goes deeper
than just remembering facts.
The AI could learn your preferences over
time. Do you like detailed explanations
or quick summaries, formal tone or
casual conversation?
If you always ask for JavaScript
examples, it might start giving those by
default.
If it knows you prefer metric units,
it'll use those without asking.
Altman actually called this long-term
memory his favorite feature in
development, and I can see why.
It's a strategic shift from raw
intelligence to building an actual
relationship with the user.
The AI stops being a smart stranger and
becomes more like a colleague who knows
you. Now, you might be wondering about
privacy, and you should be. If the AI is
remembering things about you, where is
that data stored? Is it secure? Altman
acknowledged this concern. Currently,
even temporary conversation history
isn't encrypted on OpenAI servers, which
is kind of wild. But he mentioned they
could add encryption for future models
to protect user data. So, along with
building memory into GPT6, they'll need
strong safeguards.
We'll probably see features letting you
delete or manage what the AI remembers,
balancing usefulness with privacy.
You'll want control over what gets
stored and what doesn't. Key feature
two, agentic autonomy. All right, next
big feature, and this one's a gamecher.
GPT6 is expected to have agentic
capability, which is basically a fancy
way of saying it can act more
autonomously to accomplish tasks. Up
until now, using chat GPT has been super
interactive. You ask for something, it
responds.
If you need multiple steps done, you
have to orchestrate those steps with
multiple prompts. It's like having an
assistant who needs constant direction.
But there's been this growing trend
toward autonomous AI agents, systems
that can chain together actions by
themselves to meet a goal.
OpenAI is clearly eyeing this space.
They've been working on an agents SDK,
which signals that future models will
seamlessly integrate tool use and
multi-step planning.
GPT6 is expected to lean into this hard.
So instead of just giving you an answer
or a single output, it could take
initiative and perform a series of
actions on your behalf. Let me give you
an example. Say you tell GPT6, "Help me
plan a weekend getaway." The old way,
the AI gives you a list of suggestions
and then you manually do all the
bookings. The new way with GPT6,
it could autonomously break the task
into subtasks. search for flights, find
hotels, compare prices, maybe even book
reservations with your permission, and
present you with a complete itinerary,
all with minimal handholding. Here's
another scenario. You ask for an email
draft.
An Agentic GPT6 might not only write the
draft, but also find the recipient's
email address in your contacts and cue
the email for sending, assuming it has
the necessary permissions.
If you need data analysis, it could run
code or queries in the background and
show you the results. For a complex
request like set up a website for my new
bakery, it might generate a to-do list
by domain, choose website builder,
design a logo, write content, and then
actually attempt to execute many of
those steps. This makes the AI feel less
like a chat partner and more like a
digital assistant or intern that can
carry out tasks end to end. It's moving
from being just a writer to being a
doer.
Open AAI sees huge opportunity here,
especially as rivals are exploring
similar capabilities. Of course, this
raises questions. How do we ensure the
AI doesn't go rogue? How do we govern
what it's allowed to do? How do we make
sure it checks back with users for
important confirmations?
These are active areas of research. But
the bottom line is GPT6 could save you
tons of clicks and keystrokes by
handling multi-step jobs autonomously.
It's getting closer to that sci-fi
vision of an AI you can delegate tasks
to and trust it'll get them done. Key
feature three, multimodality and other
improvements.
We've covered memory and autonomy, but
what else is GPT6 bringing to the table?
One major advancement is enhanced
multimodality.
GPT4 introduced us to multimodal AI by
accepting images as input. You could
show it a picture and it could describe
or analyze it. GPT5 built on that with
voice and better file handling. GPT6 is
expected to take it even further. We're
talking native support for not just text
and images, but audio and even video in
a more integrated way. Imagine feeding
GPT6 a short video clip and asking
what's happening here and it could step
through the video describing events or
identifying issues in real time. Or
picture this. You hum a tune and GPT6
generates a complete song or musical
score around it.
These are speculative but they fit the
pattern of each GPT generation widening
its input output capabilities.
By the time we get to GPT6, interacting
with AI might feel less like traditional
chat and more like a rich multiensory
experience.
You talk to it, show it things, maybe it
shows or plays things back. Another
improvement everyone's hoping for.
Better reasoning and reliability.
Current models, even GPT4 and 5,
occasionally stumble on logic or math.
Sometimes they give answers that sound
convincing but are completely wrong.
those infamous AI hallucinations.
GPT6 is expected to reduce these
mistakes significantly.
Open AAI has been training models not
just to be knowledgeable, but to reason
through complex problems with less human
guidance.
Alman has hinted that while GPT3 and 4
were mainly about scale and pre-training
on internet data, GPT5 and GPT6 are
venturing into reinforcement learning,
letting the AI learn through trial,
experimentation, and discovery.
He described it as like discovering new
science such as new algorithms, physics,
and biology.
That's pretty abstract, but it suggests
a fundamental shift. Not just feeding
GPT6 more data, but allowing it to learn
in entirely new ways to become smarter
and more general in its capabilities.
Now, will GPT6 be dramatically bigger in
terms of parameters than GPT5?
Some rumors claim it might break the
trillion parameter mark, which would be
unprecedented.
But Alman has downplayed the just make
it bigger approach. The focus seems to
be on efficiency and quality over sheer
size. There's even talk of dynamic
scaling. GPT6 could allocate more
computing power for hard problems and
less for easy ones. For everyday
questions, it might use a lightweight
approach, so it's fast and cost
effective. But when you throw a tough,
complex task at it, it can ramp up and
bring out the big guns.
This kind of smart efficiency would be
crucial because running these huge
models is expensive and slow if not
managed cleverly. So to sum up, GPT6 is
aiming to be more personal, more
actionoriented, more multimodal, and
more logically reliable than any AI
model we've seen. OpenAI is saying, "We
have enough raw IQ. Now let's give the
AI a better memory, broader senses, a
bit of personality, and the ability to
actually do things for you.
If they pull it off, GPT6 could feel
less like a chatbot and more like a true
AI assistant that integrates into your
life and work. Timeline.
When can we expect GPT6?
The million-dollar question. When is
GPT6 actually coming out? Here's what we
know for certain. OpenAI confirmed GPT6
won't be released in 2025,
but Altman has strongly hinted, "We
won't wait as long as we did between
GPT4 and GPT5."
That was a 2.5-year gap. He's committed
internally to faster upgrade cycles,
saying they rarely have targets further
than 6 months out. Current speculation
points to 2026, with mid 2026 being the
most likely release window.
Some analysts think early 2026 is
possible if development goes smoothly.
What's backing this up? Development is
already underway. Back in March 2024, a
former Google engineer leaked that
Microsoft was provisioning around
100,000 Nvidia H100 GPUs for GPT6
training. So massive, they said it could
affect regional power grids.
By mid 2025, OpenAI announced they were
developing over 5 gawatt of data center
capacity and had started early training
workloads.
Altman's been talking about GPT6 since
August 2025, right after GPT5's launch.
Some believe training began as early as
mid 2025, which makes late 2026
plausible after testing and safety
checks.
And those safety checks matter. Open AAI
won't release GPT6 until they're
confident it's aligned and won't behave
dangerously. So bottom line, late 2026
is a safe bet for full public rollout
with mid 2026 as an optimistic scenario.
We'll likely hear official news or a
teaser in the first half of 2026 if
progress stays on track.
How will GPT6 launch
access for users?
How will GPT6 actually reach you? Based
on OpenAI's past patterns, here's what
to expect. Initial launch will be for
paid chat GPT users first. GPT6 will
almost certainly start with chat GPT
plus 20 month or pro 200 month
subscribers. If it's computationally
intensive, they might introduce a new
pricing tier or make it pro only
initially.
US users typically get access first. So
if you're states side with a paid
subscription, you'll be first in line.
Developers will get API access, but
likely on a limited basis at first,
probably a private beta for select
partners, then broader availability.
Enterprise partners like Microsoft's
Azure OpenAI service could get access
around the same time or even before the
public. Don't be surprised if Microsoft
announces GPT6 powering new Windows or
Office features on launch day. Free chat
GPT users. Historically, the free tier
gets new models much later, if ever.
GPT4 only recently became partially
available to free users. You'll likely
need a paid subscription to access GPT6
for quite a while. Expect a cautious
roll out labeled as beta or preview
initially. Some features like full
long-term memory or advanced autonomy
might be slowly enabled as Open AI
monitors usage and ensures safety.
Practical advice. Stay tuned to OpenAI's
blog and Twitter for announcements and
consider getting on a paid plan.
Developers should watch the OpenAI forum
for API preview invitations. If all
these features pan out, GPT6 could mark
a real turning point in how we interact
with AI. We're moving from sessionbound
chat bots to an AI that becomes a
persistent personal assistant.
It's the difference between having a
smart tool you use and having a smart
partner that works with you
continuously.
GPT6's long-term memory could make
conversations feel like an ongoing
dialogue with a colleague who knows you.
Its agentic abilities might mean less
clicking, more delegating. You could say
GPT6 handle this and trust it to
execute. and its improved multimodal
understanding will broaden what we can
accomplish from creative projects to
daily tasks.
Of course, with these advancements come
responsibilities.
Open AI will need strong privacy
safeguards, misuse prevention, and
alignment to ensure the AI's output stay
grounded in reality and human values.
Alman's team is working with
psychologists on emotional AI connection
and being vocal about avoiding harm.
There are skeptics GPT5 didn't blow
everyone's mind. So, will GPT6 truly
leap ahead? Some think we're hitting a
plateau where gains get smaller. It'll
be fascinating to see if GPT6 delivers
that wow factor or mostly refineses what
exists. But one thing's certain, OpenAI
isn't just making a bigger model.
They're rethinking what AI companions
should do for us.
Altman's goal is to make Chat GPT truly
personal and revolutionary.
If GPT6 succeeds, we might look back on
it as the model that transformed AI from
a clever oracle into a trusted,
proactive partner. If you found this
helpful, hit subscribe and drop a
comment with your GPT6 predictions. What
features excite you? What concerns do
you have? Thanks for watching and I'll
see you in the next