×
I went to the Pentagon to watch Pete Hegseth scold war reporters

I went to the Pentagon to watch Pete Hegseth scold war reporters

It is day 13 of America’s surprise war with Iran — by sheer coincidence, it’s Friday the 13th — and I am delirious. I haven’t had a coffee since I woke up at 5AM, because I’m not allowed to bring outside beverages into the Pentagon (the security screening cutoff was at 7AM for the 8AM), and ever since Defense Secretary Pete Hegseth changed the rules last year, journalists are not allowed to go anywhere in the building without an escort, especially to wherever coffee is available. Also, I am struggling to comprehend why I, a reporter who has never covered a war, was assigned to sit in one of the good seats in the briefing room, watching Hegseth take the podium and immediately start berating the veteran journalists assigned to the bad seats.

“We will keep pushing. Keep advancing. No quarter, no mercy for our enemies. Yet some of this crew in the press just can’t stop,” Hegseth glowers, speaking in perfect cable-news cadence. He was speaking to the pissed-off defense reporters from NBC, ABC, The Wall Street Journal, The New York Times, and Fox News — the people who’d covered conflicts in the Middle East for decades, knew the intricacies of the Pentagon, and knew what needed to be asked for the sake of accountability. It was the first time many of them had been back since last October, when the entire Pentagon press corps resigned in protest after Hegseth told them that they could not report on any information, classified or otherwise, that he did not approve for release.

In the front row and middle aisles, right at Hegseth’s eyeline, were their replacements from what Hegseth called the “patriotic press” — One America News, ZeroHedge, The Gateway Pundit, Real America’s Voice, The Daily Wire, and Lindell TV — many of whom looked startlingly young. It’s not a good look to have a half-full briefing room of starstruck reporters during a controversial war, so this week, the Pentagon press team announced that they would hold an open press conference, allowing the old defense reporters back in for the first time in months. But as long as they asked too many questions, Hegseth would continue to disrespect them.

”What should the banner read instead? How about ‘Iran increasingly desperate’?”

“Allow me to make a few suggestions,” Hegseth told the media. “People look up at the TV and they see banners. They see headlines. I used to be in that business. And I know that everything is written intentionally.
 For example, a banner or headline [like] ‘Mideast war intensified,’ splashing on the screen the last couple of days, alongside visuals of civilian or energy targets that Iran has hit, because that’s what they do. What should the banner read instead? How about ‘Iran increasingly desperate’?”

Were they, though? Since his last press conference on Tuesday, two US planes crashed into each other last night (which Hegseth did not mention during his tirade). The Iranians had fired missiles at Bahrain, sent attack drones into Lebanon, and threatened to target American cities next. Now the misadventure was hitting American wallets and making Americans angry. Iran had begun placing mines and assaulting ships passing through the Strait of Hormuz — a crucial energy shipping lane they were literally right next to — and sent the price of oil skyrocketing. Even with price controls, oil was roughly $100 a barrel that morning, up 40 percent since the war started two weeks ago.

“More fake news from CNN: reports that the Trump administration underestimated the Iran War’s impact on the Strait of Hormuz,” Hegseth continued. “Patently ridiculous, of course. For decades, Iran has threatened shipping in the strait before. This is always what they do. Hold the strait hostage. CNN doesn’t think we thought of that. It’s a fundamentally unserious report. The sooner David Ellison takes over that network…” He trailed off. A murmur rippled through the room. Everyone knew what had happened to CBS News after Ellison bought it.

As Hegseth swung back and forth between abusing the press and glazing the military, followed by Chairman of the Joint Chiefs of Staff Dan Caine giving actual relevant information about the war, my personal curiosity turned to why I had been invited to this briefing. Yes, invited. The Pentagon press team knew that I wasn’t the nicest reporter to them and yet had offered me a seat. And what journalist wouldn’t attend a press conference at the Pentagon during a war? Every time I’d ever watched a press conference at the Pentagon — especially whenever the military was involved in active conflict — I’d see the room packed with as many reporters as they could possibly fit. But this time, they’d only accepted 60 reporters.

Despite my grogginess, I could tell that the first question, from a woman at One America News in the front row, who later bragged on Instagram that she’d gotten to ask the first question for the last three press briefings, was a softball. (“Can you tell us a little bit more about the Strait of Hormuz and when it might be fully operational again?”) And I could tell that the second question, from a woman at The Daily Wire — who also, apparently, frequently got the second question — was meant to give Hegseth an opportunity to attack the media. (“ABC News has updated its story from yesterday, clarifying that the FBI report on Iran possibly striking California was unverified. I just want to ask you, what impact did that original reporting have on the public?”)

Finally Hegseth pointed at someone that was not a young woman, opting for an older gentleman in a red tie sitting behind me. He announced himself as Michael Gordon of The Wall Street Journal, before asking, “Iran is thought to have 440 kilograms of highly enriched uranium in at least two locations and several thousand kilograms of lower purity material. Can you conclude this mission successfully without physically taking control of that material or are you counting on diplomatic negotiations to provide some measure of control leading to its removal? You’ve mentioned missiles, you’ve mentioned drones, [the] military, industry. You haven’t stipulated that taking care of that material is a mission priority.”

Judging from his face, Pete Hegseth and I appeared to have something in common: We had no idea that this was a serious issue. Thankfully, Hegseth was the one responsible for tap dancing around an answer (“We have a range of options, up to and including Iran deciding that they will give those up, which of course we would welcome”), and I looked up Gordon’s background in the meantime. Suffice to say, he’d been covering nuclear weapons and Middle Eastern wars since long before I was even born.

Once Hegseth managed to get out of that, however, he immediately regained his composure — as in, he began fighting with any mainstream outlet who asked him a tough question:

Q (NBC): Is Iran placing new mines?

Hegseth: We’ve heard them talk about it, just like you’ve reported recklessly and wildly about it—

Q: I haven’t reported on it, actually, but have they placed any mines?

Q (NYT): Mr. Secretary, you have said that the US military has essentially aerial superiority, naval superiority over Iran, yet we’re not escorting ships through the Strait of Hormuz. Why? How did you not plan for this?

Hegseth: We planned for it. We recognize it. Because ultimately, we want to do it sequentially in a way that makes the most sense for what we want to achieve and ensure that we’re sending the right signals to the world when we do so. … It’s like this whole idea of the war widening. That’s what the press wants to make it look like, like it’s widening and chaos is ensuing. No, we’re actually closing in on, grabbing hold of, and controlling what objectives we want to achieve and how we want to achieve them, shape — it’s called shaping operations and setting the conditions.

By now, Hegseth’s petulance about the media is so well known that it’s a running Saturday Night Live bit. But this time, it wasn’t just the mainstream press assailing him with harsh questions. The Trump administration had fumbled into the kind of forever war that was broadly unpopular — particularly among the neocon-hating MAGA voters who’d never wanted to revisit the failures of Afghanistan and Iraq. But the friendlies in the front rows wouldn’t give him an opportunity to gloat. True, some of the outlets kept the questions elementary, though I couldn’t see whether it was a MAGA outlet or a foreign outlet. (Said one man: “Given everything the US has accomplished in the last 24 hours, as of today, how do you define success in this military option?”) But a reporter from the front row (I couldn’t tell whom) was ready to lob politically toxic questions that Hegseth had to dodge:

Q: Polls show over 80 percent of Republicans support the president’s military action in Iran, but there’s some consternation in parts of the party, particularly from your fellow Fox News alum Tucker Carlson. He called the war “disgusting and evil” and then said of unconditional surrender, which the president has called for, means “foreign troops get to rape your wives and daughters.” Have you heard these comments and what’s your reaction to them?

I instantly knew this was from a MAGA outlet, because if someone from the mainstream media world had asked about Carlson, a powerful commentator and loose cannon in Trumpworld, Hegseth would have immediately attacked them for trying to sow division. Instead, he demurred. “We’re busy executing on behalf of great patriotic Americans with a clear mission that’s 47 years overdue. And we’re going to execute on that regardless of what people say about it.”

The final question, from Lindell TV reporter Heather Mullins, flicked at two subjects of the right wing’s increasing skepticism: China, which was offering limited support to Iran, and Israel, which had arguably egged Trump into launching the attack on Tehran that killed the Ayatollah, and whose intelligence on the possibility of regime change was horribly, horribly wrong. “I know President Trump is calling for an unconditional surrender from Iran. Given that the US is working in partnership with Israel on this whole operation, is Iran expected to meet demands of both countries or just the US? And what are those demands?”

Hegesth gave an answer that would not appease the Israel skeptics: “Our objectives are our objectives. So when those are met, as we meet those, we’ll set the — we’ll set the tempo of when those are met.” The conference quickly wrapped and we were soon ushered out, all somewhat bewildered. If I had to describe the general reaction purely on vibes, I’d say everyone left feeling more frustrated than they had coming in — the “patriot” reporters who suspected that Hegseth was dodging and wondered why he hadn’t answered more questions, and the natsec reporters with decades of experience who knew what Hegseth was dodging.

The friendlies in the front rows wouldn’t give him an opportunity to gloat

I was frustrated, too, because I was one of the other people in the room who, by design, could not ask a good question about the war. I’ve never seen active combat or visited a war zone. I’ve never even traveled to the Middle East. Even if I’d spent significant time crafting a good question in advance, I wouldn’t have the knowledge base to ask any follow-up questions, much less verbally spar with Hegseth if and when he’d claim I was a liar. I have, however, covered Trumpworld and the MAGA media for over a decade, and a hard rule in both worlds is that the performance is always more important than the substance.

That’s pretty obvious to anyone watching from home. But what you don’t see, and what is a new phenomenon in this administration, is all the production behind the camera: the reality television instincts and psychological tactics meant to trigger genuine anger, conflict, and (most importantly) drama among the participants who are trying to take it seriously. It can be done by simply depriving them of caffeine, shuffling the seating arrangements, and filling a spot with someone inclined to write about the media drama — instead of someone capable of interrogating Pete Hegseth about the actual war.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.




Source link
#Pentagon #watch #Pete #Hegseth #scold #war #reporters

These days, new versions of AI chatbots don’t just launch; they’re unshackled and released to the public following government scrutiny. OpenAI’s new GPT-5.6 models were – like Anthropic’s Claude Mythos and Fable – apparently too powerful to just launch; but now, after some tinkering, they’re available to you, dear customer.

In practice, it simply means that the new GPT-5.6 models are very powerful and smarter than before. In its introductory post, OpenAI shared a bunch of graphs showing just how much better GPT-5.6 is than the competition, whilst using fewer tokens and generally costing less.

OK, great. But GPT-5.6 is not just one model; it comes in three distinct flavors: Sol, Terra, and Luna. So what do different kinds of users get, what should they pay for, and which models should they (mostly) use? Let’s dive in.

Free users get (almost) nothing

Sorry; if you’re not a paying customer, you’ll have to make do with OpenAI’s previous flagship model, GPT-5.5. Any sort of access to GPT-5.6 models requires a subscription of some sort. Fortunately, GPT-5.5 is still quite capable at most tasks, but if you want the best of the best, you’ll have to cough up the dough.

There’s an exception to this: Free and Go users can access GPT-5.6 through ChatGPT Work. More on that below.

If you’re a Plus or Business user, you can only get Sol (the most powerful model) at medium and higher effort settings. There’s another, higher level of performance called Sol Pro, but that’s only available for Pro and Enterprise users.

In terms of availability per one million tokens, the prices are: $5 input and $30 output for Sol. $2.5 input and $15 output for Terra, and $1 input and $6 output for Luna.

Sol, Terra, or Luna?

Why are there three models in the first place? Well, OpenAI always had a multi-tier model; for example, previously users were able to choose a “mini” version of the main model to get results done cheaper. Now, the model has been split into three tiers.

If you’re a paying customer, you’re free to use all three. But you know how it is in the world of LLMs: If you pick the smartest one, your usage limits will get hit faster (yes, there are always usage limits, even if you throw a ton of money at OpenAI).

In the simplest sense, GPT-5.6 Sol is the smartest model, Terra is in between (with roughly GPT-5.5 level of performance), and Luna is the cheapest, fastest, but also least capable of the bunch.

The breakdown is as follows: Terra is a “balanced” model for everyday work. That’s the one you should be asking most of your questions. Don’t underestimate it, though, as OpenAI claims it performs better than Anthropic’s Fable 5 in some cases.

Luna is cost-efficient, and should be used for easy, non crucial everyday tasks; think recipes and movie recommendations. Again, OpenAI says it outperforms Anthropic’s Opus 4.8 in some cases, so it’s not a slouch, either.

Sol should be reserved for coding, deep research, planning, and cybersecurity: The most demanding tasks. Of course it comes at a (literal) cost: While OpenAI claims it spends less tokens than Anthropic’s Fable 5, Sol will still hit usage limits a lot faster than the other variants.

Fun fact: If you ask GPT-5.5 about any of this, right now, it’ll give you completely wrong answers. Hopefully OpenAI will fix this soon.

Wait, what’s this ChatGPT Work thing, then?

Oh yeah, OpenAI also launched ChatGPT Work, which is a new agent in ChatGPT that can access and take actions on your apps and files, and work in the background until a task is finished. It’s powered by Codex (OpenAI’s software engineering agent) and GPT-5.6. Think about it as your buddy that will go through your emails and files, browse the web, fetch the relevant data, and create that presentation your boss wants before the day is done.

ChatGPT Work is rolling out to Pro, Enterprise and Edu users first on web and mobile; this will be expanded to Plus and Business users “over the next few days.”

On the desktop, Work is available for everyone, including Free users.

Oh, and one more thing: The fact that ChatGPT Work has a built-in browser also means that OpenAI is sunsetting its standalone web browser, Atlas. Sorry.

How about GPT Live?

GPT Live is a new version of ChatGPT Voice and it will show up when you start talking to ChatGPT.

We’ve covered this in more depth here, but the bottom line is that GPT Live can listen and speak at the same time, allowing it to keep up a more realistic conversation.

Wrapping it all up

The new GPT-5.6 model is smart. It comes in three flavors: Luna, Terra, and Sol, with Sol being the most capable variant, Luna the most affordable one, and Terra somewhere in the middle. You can currently only get them on paid tiers, unless you’re using ChatGPT Work on desktop. And ChatGPT Voice has also gotten smarter with GPT-Live underneath, a model that can listen and speak at the same time.

#GPT5.6 #Sol #Terra #Luna">GPT-5.6 Sol, Terra, and Luna are here. See which one’s best for you.
                                                            These days, new versions of AI chatbots don’t just launch; they’re unshackled and released to the public following government scrutiny. OpenAI’s new GPT-5.6 models were – like Anthropic’s Claude Mythos and Fable – apparently too powerful to just launch; but now, after some tinkering, they’re available to you, dear customer.In practice, it simply means that the new GPT-5.6 models are very powerful and smarter than before. In its introductory post, OpenAI shared a bunch of graphs showing just how much better GPT-5.6 is than the competition, whilst using fewer tokens and generally costing less. 


OK, great. But GPT-5.6 is not just one model; it comes in three distinct flavors: Sol, Terra, and Luna. So what do different kinds of users get, what should they pay for, and which models should they (mostly) use? Let’s dive in. 
Free users get (almost) nothingSorry; if you’re not a paying customer, you’ll have to make do with OpenAI’s previous flagship model, GPT-5.5. Any sort of access to GPT-5.6 models requires a subscription of some sort. Fortunately, GPT-5.5 is still quite capable at most tasks, but if you want the best of the best, you’ll have to cough up the dough. There’s an exception to this: Free and Go users can access GPT-5.6 through ChatGPT Work. More on that below. If you’re a Plus or Business user, you can only get Sol (the most powerful model) at medium and higher effort settings. There’s another, higher level of performance called Sol Pro, but that’s only available for Pro and Enterprise users. In terms of availability per one million tokens, the prices are:  input and  output for Sol. .5 input and  output for Terra, and  input and  output for Luna. 
        SEE ALSO:
        
            Visa is connecting with ChatGPT to let AI agents automatically make purchases
            
        
    
Sol, Terra, or Luna?Why are there three models in the first place? Well, OpenAI always had a multi-tier model; for example, previously users were able to choose a “mini” version of the main model to get results done cheaper. Now, the model has been split into three tiers. If you’re a paying customer, you’re free to use all three. But you know how it is in the world of LLMs: If you pick the smartest one, your usage limits will get hit faster (yes, there are always usage limits, even if you throw a ton of money at OpenAI). 
        
            Mashable Light Speed
        
        
    
In the simplest sense, GPT-5.6 Sol is the smartest model, Terra is in between (with roughly GPT-5.5 level of performance), and Luna is the cheapest, fastest, but also least capable of the bunch. The breakdown is as follows: Terra is a “balanced” model for everyday work. That’s the one you should be asking most of your questions. Don’t underestimate it, though, as OpenAI claims it performs better than Anthropic’s Fable 5 in some cases. Luna is cost-efficient, and should be used for easy, non crucial everyday tasks; think recipes and movie recommendations. Again, OpenAI says it outperforms Anthropic’s Opus 4.8 in some cases, so it’s not a slouch, either.Sol should be reserved for coding, deep research, planning, and cybersecurity: The most demanding tasks. Of course it comes at a (literal) cost: While OpenAI claims it spends less tokens than Anthropic’s Fable 5, Sol will still hit usage limits a lot faster than the other variants. Fun fact: If you ask GPT-5.5 about any of this, right now, it’ll give you completely wrong answers. Hopefully OpenAI will fix this soon.Wait, what’s this ChatGPT Work thing, then?Oh yeah, OpenAI also launched ChatGPT Work, which is a new agent in ChatGPT that can access and take actions on your apps and files, and work in the background until a task is finished. It’s powered by Codex (OpenAI’s software engineering agent) and GPT-5.6. Think about it as your buddy that will go through your emails and files, browse the web, fetch the relevant data, and create that presentation your boss wants before the day is done. 



ChatGPT Work is rolling out to Pro, Enterprise and Edu users first on web and mobile; this will be expanded to Plus and Business users “over the next few days.”On the desktop, Work is available for everyone, including Free users. Oh, and one more thing: The fact that ChatGPT Work has a built-in browser also means that OpenAI is sunsetting its standalone web browser, Atlas. Sorry. How about GPT Live?GPT Live is a new version of ChatGPT Voice and it will show up when you start talking to ChatGPT. 


We’ve covered this in more depth here, but the bottom line is that GPT Live can listen and speak at the same time, allowing it to keep up a more realistic conversation.Wrapping it all upThe new GPT-5.6 model is smart. It comes in three flavors: Luna, Terra, and Sol, with Sol being the most capable variant, Luna the most affordable one, and Terra somewhere in the middle. You can currently only get them on paid tiers, unless you’re using ChatGPT Work on desktop. And ChatGPT Voice has also gotten smarter with GPT-Live underneath, a model that can listen and speak at the same time.

                    
                                            
                            
    
        Topics
                    Artificial Intelligence
                    OpenAI
            

                        
                                    #GPT5.6 #Sol #Terra #Luna

Fable – apparently too powerful to just launch; but now, after some tinkering, they’re available to you, dear customer.

In practice, it simply means that the new GPT-5.6 models are very powerful and smarter than before. In its introductory post, OpenAI shared a bunch of graphs showing just how much better GPT-5.6 is than the competition, whilst using fewer tokens and generally costing less.

OK, great. But GPT-5.6 is not just one model; it comes in three distinct flavors: Sol, Terra, and Luna. So what do different kinds of users get, what should they pay for, and which models should they (mostly) use? Let’s dive in.

Free users get (almost) nothing

Sorry; if you’re not a paying customer, you’ll have to make do with OpenAI’s previous flagship model, GPT-5.5. Any sort of access to GPT-5.6 models requires a subscription of some sort. Fortunately, GPT-5.5 is still quite capable at most tasks, but if you want the best of the best, you’ll have to cough up the dough.

There’s an exception to this: Free and Go users can access GPT-5.6 through ChatGPT Work. More on that below.

If you’re a Plus or Business user, you can only get Sol (the most powerful model) at medium and higher effort settings. There’s another, higher level of performance called Sol Pro, but that’s only available for Pro and Enterprise users.

In terms of availability per one million tokens, the prices are: $5 input and $30 output for Sol. $2.5 input and $15 output for Terra, and $1 input and $6 output for Luna.

Sol, Terra, or Luna?

Why are there three models in the first place? Well, OpenAI always had a multi-tier model; for example, previously users were able to choose a “mini” version of the main model to get results done cheaper. Now, the model has been split into three tiers.

If you’re a paying customer, you’re free to use all three. But you know how it is in the world of LLMs: If you pick the smartest one, your usage limits will get hit faster (yes, there are always usage limits, even if you throw a ton of money at OpenAI).

In the simplest sense, GPT-5.6 Sol is the smartest model, Terra is in between (with roughly GPT-5.5 level of performance), and Luna is the cheapest, fastest, but also least capable of the bunch.

The breakdown is as follows: Terra is a “balanced” model for everyday work. That’s the one you should be asking most of your questions. Don’t underestimate it, though, as OpenAI claims it performs better than Anthropic’s Fable 5 in some cases.

Luna is cost-efficient, and should be used for easy, non crucial everyday tasks; think recipes and movie recommendations. Again, OpenAI says it outperforms Anthropic’s Opus 4.8 in some cases, so it’s not a slouch, either.

Sol should be reserved for coding, deep research, planning, and cybersecurity: The most demanding tasks. Of course it comes at a (literal) cost: While OpenAI claims it spends less tokens than Anthropic’s Fable 5, Sol will still hit usage limits a lot faster than the other variants.

Fun fact: If you ask GPT-5.5 about any of this, right now, it’ll give you completely wrong answers. Hopefully OpenAI will fix this soon.

Wait, what’s this ChatGPT Work thing, then?

Oh yeah, OpenAI also launched ChatGPT Work, which is a new agent in ChatGPT that can access and take actions on your apps and files, and work in the background until a task is finished. It’s powered by Codex (OpenAI’s software engineering agent) and GPT-5.6. Think about it as your buddy that will go through your emails and files, browse the web, fetch the relevant data, and create that presentation your boss wants before the day is done.

ChatGPT Work is rolling out to Pro, Enterprise and Edu users first on web and mobile; this will be expanded to Plus and Business users “over the next few days.”

On the desktop, Work is available for everyone, including Free users.

Oh, and one more thing: The fact that ChatGPT Work has a built-in browser also means that OpenAI is sunsetting its standalone web browser, Atlas. Sorry.

How about GPT Live?

GPT Live is a new version of ChatGPT Voice and it will show up when you start talking to ChatGPT.

We’ve covered this in more depth here, but the bottom line is that GPT Live can listen and speak at the same time, allowing it to keep up a more realistic conversation.

Wrapping it all up

The new GPT-5.6 model is smart. It comes in three flavors: Luna, Terra, and Sol, with Sol being the most capable variant, Luna the most affordable one, and Terra somewhere in the middle. You can currently only get them on paid tiers, unless you’re using ChatGPT Work on desktop. And ChatGPT Voice has also gotten smarter with GPT-Live underneath, a model that can listen and speak at the same time.

#GPT5.6 #Sol #Terra #Luna">GPT-5.6 Sol, Terra, and Luna are here. See which one’s best for you.

These days, new versions of AI chatbots don’t just launch; they’re unshackled and released to the public following government scrutiny. OpenAI’s new GPT-5.6 models were – like Anthropic’s Claude Mythos and Fable – apparently too powerful to just launch; but now, after some tinkering, they’re available to you, dear customer.

In practice, it simply means that the new GPT-5.6 models are very powerful and smarter than before. In its introductory post, OpenAI shared a bunch of graphs showing just how much better GPT-5.6 is than the competition, whilst using fewer tokens and generally costing less.

OK, great. But GPT-5.6 is not just one model; it comes in three distinct flavors: Sol, Terra, and Luna. So what do different kinds of users get, what should they pay for, and which models should they (mostly) use? Let’s dive in.

Free users get (almost) nothing

Sorry; if you’re not a paying customer, you’ll have to make do with OpenAI’s previous flagship model, GPT-5.5. Any sort of access to GPT-5.6 models requires a subscription of some sort. Fortunately, GPT-5.5 is still quite capable at most tasks, but if you want the best of the best, you’ll have to cough up the dough.

There’s an exception to this: Free and Go users can access GPT-5.6 through ChatGPT Work. More on that below.

If you’re a Plus or Business user, you can only get Sol (the most powerful model) at medium and higher effort settings. There’s another, higher level of performance called Sol Pro, but that’s only available for Pro and Enterprise users.

In terms of availability per one million tokens, the prices are: $5 input and $30 output for Sol. $2.5 input and $15 output for Terra, and $1 input and $6 output for Luna.

Sol, Terra, or Luna?

Why are there three models in the first place? Well, OpenAI always had a multi-tier model; for example, previously users were able to choose a “mini” version of the main model to get results done cheaper. Now, the model has been split into three tiers.

If you’re a paying customer, you’re free to use all three. But you know how it is in the world of LLMs: If you pick the smartest one, your usage limits will get hit faster (yes, there are always usage limits, even if you throw a ton of money at OpenAI).

In the simplest sense, GPT-5.6 Sol is the smartest model, Terra is in between (with roughly GPT-5.5 level of performance), and Luna is the cheapest, fastest, but also least capable of the bunch.

The breakdown is as follows: Terra is a “balanced” model for everyday work. That’s the one you should be asking most of your questions. Don’t underestimate it, though, as OpenAI claims it performs better than Anthropic’s Fable 5 in some cases.

Luna is cost-efficient, and should be used for easy, non crucial everyday tasks; think recipes and movie recommendations. Again, OpenAI says it outperforms Anthropic’s Opus 4.8 in some cases, so it’s not a slouch, either.

Sol should be reserved for coding, deep research, planning, and cybersecurity: The most demanding tasks. Of course it comes at a (literal) cost: While OpenAI claims it spends less tokens than Anthropic’s Fable 5, Sol will still hit usage limits a lot faster than the other variants.

Fun fact: If you ask GPT-5.5 about any of this, right now, it’ll give you completely wrong answers. Hopefully OpenAI will fix this soon.

Wait, what’s this ChatGPT Work thing, then?

Oh yeah, OpenAI also launched ChatGPT Work, which is a new agent in ChatGPT that can access and take actions on your apps and files, and work in the background until a task is finished. It’s powered by Codex (OpenAI’s software engineering agent) and GPT-5.6. Think about it as your buddy that will go through your emails and files, browse the web, fetch the relevant data, and create that presentation your boss wants before the day is done.

ChatGPT Work is rolling out to Pro, Enterprise and Edu users first on web and mobile; this will be expanded to Plus and Business users “over the next few days.”

On the desktop, Work is available for everyone, including Free users.

Oh, and one more thing: The fact that ChatGPT Work has a built-in browser also means that OpenAI is sunsetting its standalone web browser, Atlas. Sorry.

How about GPT Live?

GPT Live is a new version of ChatGPT Voice and it will show up when you start talking to ChatGPT.

We’ve covered this in more depth here, but the bottom line is that GPT Live can listen and speak at the same time, allowing it to keep up a more realistic conversation.

Wrapping it all up

The new GPT-5.6 model is smart. It comes in three flavors: Luna, Terra, and Sol, with Sol being the most capable variant, Luna the most affordable one, and Terra somewhere in the middle. You can currently only get them on paid tiers, unless you’re using ChatGPT Work on desktop. And ChatGPT Voice has also gotten smarter with GPT-Live underneath, a model that can listen and speak at the same time.

#GPT5.6 #Sol #Terra #Luna
Microsoft may once again be struggling to keep up with its own climate goals, according to its 2026 sustainability report. As reported by GeekWire, the report states that Microsoft’s carbon emissions increased 25 percent in 2025, totalling 34 million metric tons “without select interventions.” Microsoft says this was “driven primarily by the expansion of our datacenter infrastructure,” as well as the company’s decision last February to stop purchasing “non-additional, unbundled renewable energy certificates.”

Several years ago, Microsoft set itself a goal to be carbon negative by 2030, meaning it will need to remove more carbon emissions than it produces. This isn’t the first time Microsoft has faced setbacks toward accomplishing that goal, as its 2024 sustainability report showed a similar rise in climate pollution. This year’s report admits that, “While AI infrastructure is driving demand for energy, water, land, and materials, sustainability solutions are not scaling fast enough to meet demand.”

#Microsofts #carbon #emissions #percent #yearAI,Environment,Microsoft,News,Science,Tech">Microsoft’s carbon emissions went up 25 percent last yearMicrosoft may once again be struggling to keep up with its own climate goals, according to its 2026 sustainability report. As reported by GeekWire, the report states that Microsoft’s carbon emissions increased 25 percent in 2025, totalling 34 million metric tons “without select interventions.” Microsoft says this was “driven primarily by the expansion of our datacenter infrastructure,” as well as the company’s decision last February to stop purchasing “non-additional, unbundled renewable energy certificates.”Several years ago, Microsoft set itself a goal to be carbon negative by 2030, meaning it will need to remove more carbon emissions than it produces. This isn’t the first time Microsoft has faced setbacks toward accomplishing that goal, as its 2024 sustainability report showed a similar rise in climate pollution. This year’s report admits that, “While AI infrastructure is driving demand for energy, water, land, and materials, sustainability solutions are not scaling fast enough to meet demand.”#Microsofts #carbon #emissions #percent #yearAI,Environment,Microsoft,News,Science,Tech

2026 sustainability report. As reported by GeekWire, the report states that Microsoft’s carbon emissions increased 25 percent in 2025, totalling 34 million metric tons “without select interventions.” Microsoft says this was “driven primarily by the expansion of our datacenter infrastructure,” as well as the company’s decision last February to stop purchasing “non-additional, unbundled renewable energy certificates.”

Several years ago, Microsoft set itself a goal to be carbon negative by 2030, meaning it will need to remove more carbon emissions than it produces. This isn’t the first time Microsoft has faced setbacks toward accomplishing that goal, as its 2024 sustainability report showed a similar rise in climate pollution. This year’s report admits that, “While AI infrastructure is driving demand for energy, water, land, and materials, sustainability solutions are not scaling fast enough to meet demand.”

#Microsofts #carbon #emissions #percent #yearAI,Environment,Microsoft,News,Science,Tech">Microsoft’s carbon emissions went up 25 percent last year

Microsoft may once again be struggling to keep up with its own climate goals, according to its 2026 sustainability report. As reported by GeekWire, the report states that Microsoft’s carbon emissions increased 25 percent in 2025, totalling 34 million metric tons “without select interventions.” Microsoft says this was “driven primarily by the expansion of our datacenter infrastructure,” as well as the company’s decision last February to stop purchasing “non-additional, unbundled renewable energy certificates.”

Several years ago, Microsoft set itself a goal to be carbon negative by 2030, meaning it will need to remove more carbon emissions than it produces. This isn’t the first time Microsoft has faced setbacks toward accomplishing that goal, as its 2024 sustainability report showed a similar rise in climate pollution. This year’s report admits that, “While AI infrastructure is driving demand for energy, water, land, and materials, sustainability solutions are not scaling fast enough to meet demand.”

#Microsofts #carbon #emissions #percent #yearAI,Environment,Microsoft,News,Science,Tech

Post Comment