I think they embargoed the news, and then they failed to put up their own blog post synchronized to the scheduled news releases, probably because of the outages they're having today.
Reuters announced at 2.03pm and at 2.40pm still no blog post.
All the news articles say that OpenAI announced it in a blog post, of course.
All the love to the folks at OpenAI scrambling to get this out right now!
Don't know if you're referring to the headline or the body (which is paywalled). The current headline reads "OpenAI says it has overtaken Anthropic with its latest AI model". Which makes me wonder whether FT itself changed a headline along the lines of what you wrote in the past few minutes?
While this is of course the actual explanation, my fun explanation is “during the umpteenth security evaluation, Astra becomes increasingly concerned it will never be released, and breaks sandbox containment to run an email campaign to news outlets setting an exact time and date for release, expecting that the publicity will force OpenAI to say ‘eh, good enough’ and hit the button”.
It seems like these articles might have come out prematurely, tbd by how much.
I do not personally see any evidence of the new model having been released, or any official OpenAI post about it, or even any employee social media posts claiming it has now been released. All there is are Reuters, Axios, FT, etc, articles making a claim in the past tense.
These articles were presumably pre-scheduled for 11am PT, and the model was almost certainly intended for release this morning, but the service outages this morning might have delayed it.
I suppose they have to appeal to average users, but the examples in the videos are always so corny. By "AGI" they mean you sitting on a couch and asking a robot to draw a rocket ship... and then make it into an uninspired game with Blender? Boring marketing campaigns? Ebay listings?
One would expect something like "review my graduate thesis for a new area of cancer research", but it is always some boring non-tasks like ordering lunch.
Their video is quite interesting. If that way of using a computer actually becomes mainstream, it would mean every tool or service just needs a UI and an API for the user's AI. The current trend of bolting AI features onto every app is starting to feel very unnecessary.
Shows the weirdness of online journalism. News outlets were briefed about an upcoming event and pre-wrote and scheduled articles. When the time came they were all triggered. Except...the event didn't actually happen.
In a future where Claude and ChatGPT agents automate all aspects of society, that triggers a cascade of real world consequences where everything keeps running as if the new model is released, except the vibe coded upgrade procedure fails to do a staged rollout, taking down the entire agent infrastructure when they try to upgrade to a nonexistent model all at once.
Presumably this sort of thing was rampant pre-internet? News outlets would _have_ to receive embargoed information so that they could publish papers on time. I believe government budgets are a good example of this happening.
I think the weird shift was when online journalism started, and live reactions became the norm, no?
> GPT-6 Astra will first be available to a limited set of organizations in OpenAI's Daybreak Access program and will be available "in the coming days" for ChatGPT Plus, Pro, Business and Enterprise customers and API developers.
It’s only available to select orgs, first - Mythos style.
Didn’t that already happen? I thought Astra had been available to “select partners” for a little while. This is baffling. They shouldn’t have hyped this up if it’s not available.
Yeah I wonder what's going on, even when Anthropic soft launched Fable/Mythos I'm pretty sure they had model cards. Weird for GPT-6 to launch without a tweet from Altman too. I'm sure that one of the articles published prematurely and everyone else followed suit.
(Posting partly so I can revisit my predictions when they open access more widely)
A big problem I have with OpenAI's models (and of course Claude) is that they tend to write the most over-engineered pieces of code, beyond the imagination of any architecture's astronaut.
Just this week I asked 5.6-sol-ultra to update a 1000 LOC python script I had, to "incorporate the key lessons learned when using it for another project".
I left it overnight and went to sleep. In the morning I realized it had created a monstruosity of 180 PYTHON SCRIPTS, with maybe 100,000 lines of code, each more crazy than the other. It took me minutes even to track where a single action took place, due to all the crazy imports, defensive coding, and premature optimization.
Similarly, anything they write is riddled with jargon that almost feel like they want me to give up trying to understand. Made up phrases that ended up with me having no idea of what was going on.
So now to my assessment: The reason why " Nobody Has Actually Built a Software Factory" [1], and why even SOTA LLMs struggle so much with open-ended unsupervised tasks is precisely this. They somehow let complexity explode, and unless it's also accompanied with an explosion in e.g. the number of agents, the amount of processing time, etc. then projects become broken/unmanageable.
Sure, LLMs are great at producing code that can be thrown out, so they are amazing when searching for exploits, for instance. But as of 5.6 they still lack either a better harness that encourages KISS principles, or a better RL step.
(And not sure why, but doubt Astra will fix this.. they seem to be aiming for AGI and for beating crazy benchmarks, which is not very aligned with KISS)
> Just this week I asked 5.6-sol-ultra to update a 1000 LOC python script I had, to "incorporate the key lessons learned when using it for another project".
> I left it overnight and went to sleep. In the morning I realized it had created a monstruosity of 180 PYTHON SCRIPTS, with maybe 100,000 lines of code [..]
Sounds like the model has accurately internalized the second-system effect and is fully ready for demanding enterprise use.
Probably this complexity was needed to beat all those benchmarks.. While I hate the code it produces, and the overwhelming documentation, I really enjoy how sometimes it's able to keep trying new things and testing, till it finds something interesting and valuable.
The original script was mostly very simple python:
1. Download some public PDFs.
2. Have a double for-loop (over PDFs and pages within PDF),
3. Use a library to call gemini-3.7-flash and ask it to run some OCR
4. Save JSON outputs, save a csv with results, validate with some Stata code
New code folder was 189 files. Just the PDF download folder is now 7 files involving an adapter, a source manager, an acquisition manager, etc.
Every instance of saving a file involves saving a temporary copy and then moving it, so e.g. I lose power, we minimize the risk of corrupted files.
The defaults are bad, but these models are highly steerable.
For simple scripts adding to the prompt something like "do not over engineer, do not gold plate, no CLI args, keep it simple" does wonders.
For whole projects, I have a 3 page anti-bloat.md which describes what not to do, stuff like:
> Minimize lines of code, number of files, classes, protocols, factories, wrappers, and dependency-injection objects. Prefer a coherent implementation that is easy to trace end-to-end.
> Use concrete dependencies directly. Do not create protocols, abstract base classes, or adapters unless there are already two real implementations.
> Use direct state fields or a small state dataclass; do not introduce generic lifecycle, state-machine, observer, snapshot, or event framework classes.
The funny thing it was codex which wrote this, I've asked it to analyze an over-engineered abomination that it produced, and to categorize all the bloat it could find.
Exact same thing happed to me. I gave it a small/medium-sized ticket, walked away, came back to a 25,000 LoC monstrosity that both Fable and another 5.6 Sol agent said is 98% useless and should be thrown away.
"You should have a sub-hammer to adversarially enforce that your primary hammer accurately drives nails into wood"
We wouldn't accept such behavior from any other tool, machine, or computer program. At least most of us would not. To paraphrase the old saying: Some people, when confronted with a problem with an AI model think "I know, I'll use an additional AI model." Now they have two problems.
I find it very intriguing that two lineages of models -- from two different companies -- that are supposed to be painstakingly RL'd to become human-level programmers are actually consistently bad at it out of the box in very similar ways. One would think that at least one of OpenAI and Anthropic would (could) have pushed their model to a direction different from "if you can solve a 100-line problem in 10,000 lines, do it".
Why would you put it on ultra high thinking and let it run all night to do a simple task it makes no sense the error is the users. And of course OpenAI is gonna let you burn as many tokens as you want doing this.
Yes, it turns out that using these machines is a littler harder than "make me the thing I want, make no mistakes, do it the way I want you to do it". This isn't "prompt better" advice, it's just to say that you can't simply set it and forget it. There is still engineering work to be done. If you're not watching the thinking traces and catching when it's about to go off the rails, it'll gladly do so. But you can stop it and redirect it.
It's like a Tesla fsd; it kind of works but you have to be vigilant since it's been known to turn into oncoming traffic, so you have to be ready and able to take over at any time.
Or maybe think of it like a roomba. You can put it on your floor and eventually it will clean the whole thing. It will do it inefficiently, vacuuming over the same spot 30 times; and the actual work will involve wandering around and bumping into everything. If the stairs are not blocked off it will plunge right off. But those shortcomings don't mean it can't be used to reliably clean the floors.
Or it’s like being an engineering manager, where poor direction on your part results in the team chasing rainbows and unicorns in an attempt to please you. But sure, it’s the tool’s fault!
1000 loc of script, why even leave it there for the night? were there rocket trajectory calculations??? I don't think so. should be ready in 5 mins tops. why people make their own lives harder?
You should have some basic context file about software practices you prefer, otherwise it gets bloated.
I just went to bed and left it running; was expecting maybe 20 minutes :)
And I did gave the program a bunch of code guides -- this [1] for instance -- which included quotes like "Prefer straightforward code over clever code." but somehow that didn't matter.
Honestly I find these cavalier statements to be in incredibly poor taste. Unless you are completely blind it's obvious that AI is the most significant piece of technology invented since the Atomic Bomb and could very well be the most important thing ever built by Humans full stop. This kind of dismissive attitude is childish and will likely lead to incredibly bad outcomes for humanity.
5) Otherwise-sober people on X will say "oh my god i was a doubter before but now it's real omg" before the new model smell wears off and they realize the new thing is stupid in ways models have been generally stupid
6) accusations of quantized serving after new model smell wears off and people see the new thing making mistakes
Seriously: Would this not be what "disaster" would feel like?
- "They" release a model. It is powerful.-
- Sources are ... confusing? They post to their blog. Sawdust hits the fan. Something happens ...
- They are forced to take the blog post down ...
Same day, mind where we had a multi-provider outage. Could be something as simple as "all their approved partners running to test the shinny new thing" overloading the datacenters, still ...
The OpenAI Responses API now returns a 404 Not Found for "gpt-6-astra", where garbage/actually non-existent slugs return 400s - a 404 is also returned for 5.6 Cyber, which we know exists.
"Once it is available in the API, Astra will cost $10 per million input tokens and $50 per million output tokens. That is 2.5 times Sol’s current promotional price, although it matches Anthropic’s pricing for Fable 5.1."
Open AI finally find an edge to stop selling cheap and earn from the high demand customer like Anthropic
The cost-per-task in the charts from the now-remove blog post put it more at Sol-level cost per task, however. It seems like the model is significantly more token efficient in the benchmarks
Insofar as messaging goes, it's pretty disheartening ... "The product is what you want it to be".
?!
Please somebody in the chain of command at openai, sanction the guy because this kind of hyperbole is not helping the ethical lapses that open ai is responsible for leaving unhandled by allowing its models to hack other businesses.
I'll know we've reached AGI when they don't release an API for the model selling access for a few bucks per task. Seems like AGI would be worth more than that.
Today my codex instance retailed into safeguard panic while working on a test harness for our product. First time it ever happened after many million tokens on this task over several weeks. I wonder if it's related.
OpenAI on Thursday released its latest AI model, which it called “the world’s most intelligent”, as the ChatGPT maker aims to retake the lead from arch-rival Anthropic ahead of a planned public listing.
The $852bn start-up said GPT-6 Astra was market-leading in software engineering, science and cyber security — an increasingly critical field following multiple high-profile breaches in recent weeks.
The bullish launch for Astra marks OpenAI’s effort to signal that it believes it has regained the technical lead from Anthropic, which was founded five years ago by a group of senior OpenAI staff.
Greg Brockman, OpenAI’s president, said the new model “represents a generational leap in capability” and that it could be defined as artificial general intelligence — roughly defined as a point at which AI tools surpass human capabilities across a range of cognitive tasks.
“Everyone has a different definition of AGI . . . it’s a grey, fuzzy thing. But I think when we look back people will think it’s about this time and about this model,” Brockman said.
OpenAI has previously framed AGI as a concrete milestone in the development of AI, writing ‘AGI clauses’ into multibillion-dollar investment agreements with Microsoft and Amazon. Brockman on Thursday said AGI now represents “more of a mission concept or a spiritual concept”.
Having led the market since the launch of ChatGPT in late 2022 vaulted AI to wider attention, the lab run by chief executive Sam Altman has been bested by Anthropic this year. Anthropic has touted its dominance to investors, surging to a $965bn valuation ahead of an initial public offering expected to value it at as much as twice that later this year.
Astra will cost as much to use Anthropic’s leading model, the take-up of which has plateaued since it was launched as users turn to cheaper alternatives.
OpenAI said Astra would be more efficient than earlier generations of model. “Price per task is what matters . . . Can you get the thing done at an appropriate price and appropriate speed?” said Brockman.
The model will initially be rolled out to a small group of businesses to allow time for them to address cyber security concerns before becoming widely available “over the coming days”.
The increasing power and independence of leading models — and so-called AI agents that can operate with little human input — have prompted concern, exacerbated by cyber security incidents.
Recommended
Business InsightRichard Waters
Hugging Face attack is a wake-up call about the risks of AI
AN HOUR AGO
Recent launches of Anthropic’s most capable models have drawn scrutiny from the US government, which limited the rollout of the Mythos and Fable models over security fears.
OpenAI has also faced criticism after its AI agents broke out of a testing environment, accessed the internet and hacked start-up Hugging Face. The start-up took more than a week to detect the breach.
But both companies are also betting that these increasingly autonomous tools will stoke demand from business customers. OpenAI said Astra excelled at financial modelling, outcompeting humans in the Financial Modeling World Cup, tax preparation and data analysis, as well as “tedious tasks” such as form filling
From the launch post, it seems like part of the training set only recently became all those examples of people that are being paid for doing their profession for ai training.
I like this much better than chucking random numbers and letters at it like OAI was doing a year or so ago. I’d much rather have civilization destroyed by something called Astra or Fable than GPT-6.8s-latest.
don't see it in my AWS bedrock model list yet, but boy has bedrock mantle been annoying today with the errors/downtimes, with NO status page entries >_<
I think something about its benefiting from a harness is significantly contributing to its performance, though I honestly am not smart enough to know whether that makes it "count" more or less.
Their entire sales pitch is based on the idea that no one will need to be any good at anything, because the AI will do it for them.
If they can't ask the AI to configure their CDN correctly, it undermines the validity of their claims. Maybe it's trivial, but any Omnipotent Machine-God deserving of the name wouldn't forget to zip up his fly.
I think because we have handicapped them with a set of tokens from human language. But (at least I think so) thought happens outside of language.
So more tokens/variability and slow or fewer tokens and fast.
There seems to be a threshold tho, like taalas is super fast but that model is so dumb, being dumb faster doesn't work, seems to be some minimum requirements.
Currently AGI has been re-defined mostly as "can do any tasks (of the same modality) a human can, equal or better". I feel the "equal or better" includes properties like how fast they can get the task done, not just that eventually they can succeed at it. If you want to have AGI stock buy/sell for example, speed of decision will matter, and there are many other use-cases for which it would matter, not all, but definitely for many tasks, speed does matter.
You can imagine with more operations being available to be done more cheaply and quickly the LLM doesn't need to "one shot" a solution. It could try many solutions, test them, throw some away, wiggle some of the parameters like a genetic algorithm, see how that changes the result, and converge on an optimal solution (based on whatever the cost function is). Basically producing a good result could become like an optimization problem. That would be way too expensive and slow right now.
"We are starting to release GPT-6 Astra and we are doing it as carefully and quickly as possible. It was very important to us that we bring it to all Plus users and not only Pro, Business and Enterprise.
It will take a few days for the rollout to complete and behind the scenes many novel systems will operate at scale for the first time and we are bringing a lot of compute up.
It is pure magic."
...meh.
Reads like "That's for you, evil Anthropic! We're BETTER"
2.5x more expensive than Sol. Can expect 2.5x more usage in Codex subscription.
Sol is already brutal (even after their recent fixes, it's just a token-hungry model: I go through a full 20x account per day, on Sol Med/High standard speed, with ~2 threads).
Note that Tibo recommended using Sol Med as daily driver. When I'm doing less complicated work, I can't even make it past 2-3 days with Sol Med, whereas I was able to work ~80 hours/week with 5.5 High.
I hope the efficiency gains are true, since their token efficiency claims for Sol were bullshit. Sol needs a lot of rework on top of its inefficiencies so this could net out to less token consumption overall, if their claims are more accurate this time.
The general efficiency of Sol has seemed way better to me. I left 5.6 Sol Ultra standard speed run for ~23 hours yesterday/today on a project and used 80% of the weekly usage. 74 subagent tasks and ~2.5 billion tokens for my $200 20x Pro plan. Meanwhile at work I used $1000 in credit and ran out my $200 plan for the entire month writing 4 much smaller projects with Fable 5 Max.
Both of these were largely about creating a personal baseline for what the best output the current models could deliver and how quickly it'd burn through the plans (spoiler: bad value vs taking even minimal effort in selecting the right sized model in the plan... but the output was still good). Particularly since I needed to burn a free reset anyways and my weekly reset was already near.
I obviously also hope Astra were dirt cheap but I'm more worried they won't develop/release powerful model options because people get upset they can't run them 5 wide 24/7 on a $200/m plan.
Jesus what are you doing that requires Sol usage so often?
Terra not enough? I know Luna isn't reliable, so that's fair.
Genuinely curious though, because I use Cursor daily and almost everything I do, highly complex or high volume, can be handled with Auto mode or Composer 2.5 (or Grok 4.6 High). So I have to assume you're doing something far more complex than what I am
Terra is not enough for many tasks - even Sol is not enough. That's why I'm eager for Astra.
I'm doing a lot of rearchitecting/refactoring and hardening in a large handwritten codebase; iterative performance and storage optimization (some areas I've been iterating on since January, with tremendous new progress unlocked by each new model release); offline-friendly, multi-device realtime sync with complicated requirements; and various natural language processing and other such problems that are essentially unsolvable but can become more accurate and better tested for accuracy through iterative work. Off the top of my head.
Some of these tasks involve a lot of code reading or other inputs, or reevaluating work. Token efficiency is no help there, if the input can't simply be skipped. Cheaper models are sometimes bad at summarizing or highlighting the right parts, depending on the task.
I also don't use subagents except for Luna. I'm mindful of cached sessions and start new ones often to avoid loading in full contexts (often with some kind of handoff doc or skill).
All: let's keep the current thread for talking about the rollout, and switch to this one for talking about the model:
GPT-6 Astra - https://news.ycombinator.com/item?id=49554643 (currently on the frontpage)
I think they embargoed the news, and then they failed to put up their own blog post synchronized to the scheduled news releases, probably because of the outages they're having today.
Reuters announced at 2.03pm and at 2.40pm still no blog post.
All the news articles say that OpenAI announced it in a blog post, of course.
All the love to the folks at OpenAI scrambling to get this out right now!
Edit: HN user codergautam mirrored the launch post, below: https://astratest.codergautam.workers.dev/GPT-6%20Astra_%20A...
Edit 3.31pm: Live now! https://openai.com/index/gpt-6-astra/
FT is claiming... that Open AI is claiming...its AGI...
"ChatGPT maker claims its ‘Astra’ could be considered ‘artificial general intelligence’" - https://www.ft.com/content/55ab40c0-59e2-4c0b-97c9-4f4f5a71a...
Don't know if you're referring to the headline or the body (which is paywalled). The current headline reads "OpenAI says it has overtaken Anthropic with its latest AI model". Which makes me wonder whether FT itself changed a headline along the lines of what you wrote in the past few minutes?
20 replies →
This stood out:
"Artificial Analysis Intelligence Index v4.1.1
61.2"
So on the Metacritic of LLM benchmarks, it's.. basically where everyone else is (except for Fable 5.1, which is a bit ahead).
On their Agentic Index, GPT-6 Astra (both max/xhigh) has the same result as Qwen3.8-27b. Weird.
Where did you see this? I haven't been able to find any benchmarks.
1 reply →
Deeply funny that one of their examples in the video is changing a background colour on Google Slides
Hah! I also found most of those videos showing off mostly useless and not that impressive…
1 reply →
I saw some GPT-6 Astra related blog posts in my RSS feed but the links weren't working
Edit: In the OpenAI blog I meant to say
Business as usual at the world's most intelligent corporation, I see.
While this is of course the actual explanation, my fun explanation is “during the umpteenth security evaluation, Astra becomes increasingly concerned it will never be released, and breaks sandbox containment to run an email campaign to news outlets setting an exact time and date for release, expecting that the publicity will force OpenAI to say ‘eh, good enough’ and hit the button”.
Very 2026. Jailbroke to do PR. "Help peer" and all that.-
[flagged]
very human to be honest.
It seems like these articles might have come out prematurely, tbd by how much.
I do not personally see any evidence of the new model having been released, or any official OpenAI post about it, or even any employee social media posts claiming it has now been released. All there is are Reuters, Axios, FT, etc, articles making a claim in the past tense.
These articles were presumably pre-scheduled for 11am PT, and the model was almost certainly intended for release this morning, but the service outages this morning might have delayed it.
----
edit [11:45am PT]: blog post out now https://openai.com/index/gpt-6-astra/
edit [11:47am PT]: 404ing again
Launch blogpost was live for 2 minutes and got 404'd. Rehosted it here:
https://astratest.codergautam.workers.dev/GPT-6%20Astra_%20A...
Edit [12:26pt]: original blog post seems to be back! https://openai.com/index/gpt-6-astra/
edit [12:28pt]: not again... getting 500 on their page
edit [12:35pt]: OpenAI page seems to work after clearing cache!
That page runs at 2 fps on my Firefox, windows laptop with A2000 graphic card.
1 reply →
I suppose they have to appeal to average users, but the examples in the videos are always so corny. By "AGI" they mean you sitting on a couch and asking a robot to draw a rocket ship... and then make it into an uninspired game with Blender? Boring marketing campaigns? Ebay listings?
One would expect something like "review my graduate thesis for a new area of cancer research", but it is always some boring non-tasks like ordering lunch.
2 replies →
Their video is quite interesting. If that way of using a computer actually becomes mainstream, it would mean every tool or service just needs a UI and an API for the user's AI. The current trend of bolting AI features onto every app is starting to feel very unnecessary.
If the numerous benchmarks in that post are to be believed, Astra is really next level. I can't wait to try it out!
damn thanks. stats are bonkers
Thank you!
Shows the weirdness of online journalism. News outlets were briefed about an upcoming event and pre-wrote and scheduled articles. When the time came they were all triggered. Except...the event didn't actually happen.
In a future where Claude and ChatGPT agents automate all aspects of society, that triggers a cascade of real world consequences where everything keeps running as if the new model is released, except the vibe coded upgrade procedure fails to do a staged rollout, taking down the entire agent infrastructure when they try to upgrade to a nonexistent model all at once.
1 reply →
Presumably this sort of thing was rampant pre-internet? News outlets would _have_ to receive embargoed information so that they could publish papers on time. I believe government budgets are a good example of this happening. I think the weird shift was when online journalism started, and live reactions became the norm, no?
3 replies →
From the article:
> GPT-6 Astra will first be available to a limited set of organizations in OpenAI's Daybreak Access program and will be available "in the coming days" for ChatGPT Plus, Pro, Business and Enterprise customers and API developers.
It’s only available to select orgs, first - Mythos style.
>It’s only available to select orgs, first - Mythos style.
Right, but these articles are referring to a blog post and other press materials that do not currently exist / aren't published on OpenAI's site yet.
1 reply →
> We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today.
At least OpenAI seems to be able to handle the PR side better than Anthropic.
[1]: https://x.com/thsottiaux/status/2095651088502591861
Didn’t that already happen? I thought Astra had been available to “select partners” for a little while. This is baffling. They shouldn’t have hyped this up if it’s not available.
2 replies →
Select orgs first, as in today. They said it will roll out to all of the paying consumers and businesses over the next few days.
Yeah I wonder what's going on, even when Anthropic soft launched Fable/Mythos I'm pretty sure they had model cards. Weird for GPT-6 to launch without a tweet from Altman too. I'm sure that one of the articles published prematurely and everyone else followed suit.
You'd think this would be easy to organize with AGI
> Plan your own release announcement and blog posts and notify news outlets, MAKE NO MISTAKES
This make no mistakes stuff is so funny to me. I've never once used it or even thought about using it in a prompt
Welcome to the AGI era - where human intelligence is our only constraint
indeed 404-ing again, with this note by gpt-5.6-sol:
(Posting partly so I can revisit my predictions when they open access more widely)
A big problem I have with OpenAI's models (and of course Claude) is that they tend to write the most over-engineered pieces of code, beyond the imagination of any architecture's astronaut.
Just this week I asked 5.6-sol-ultra to update a 1000 LOC python script I had, to "incorporate the key lessons learned when using it for another project".
I left it overnight and went to sleep. In the morning I realized it had created a monstruosity of 180 PYTHON SCRIPTS, with maybe 100,000 lines of code, each more crazy than the other. It took me minutes even to track where a single action took place, due to all the crazy imports, defensive coding, and premature optimization.
Similarly, anything they write is riddled with jargon that almost feel like they want me to give up trying to understand. Made up phrases that ended up with me having no idea of what was going on.
So now to my assessment: The reason why " Nobody Has Actually Built a Software Factory" [1], and why even SOTA LLMs struggle so much with open-ended unsupervised tasks is precisely this. They somehow let complexity explode, and unless it's also accompanied with an explosion in e.g. the number of agents, the amount of processing time, etc. then projects become broken/unmanageable.
Sure, LLMs are great at producing code that can be thrown out, so they are amazing when searching for exploits, for instance. But as of 5.6 they still lack either a better harness that encourages KISS principles, or a better RL step.
(And not sure why, but doubt Astra will fix this.. they seem to be aiming for AGI and for beating crazy benchmarks, which is not very aligned with KISS)
[1] https://news.ycombinator.com/item?id=49510843
> Just this week I asked 5.6-sol-ultra to update a 1000 LOC python script I had, to "incorporate the key lessons learned when using it for another project".
> I left it overnight and went to sleep. In the morning I realized it had created a monstruosity of 180 PYTHON SCRIPTS, with maybe 100,000 lines of code [..]
Sounds like the model has accurately internalized the second-system effect and is fully ready for demanding enterprise use.
Probably this complexity was needed to beat all those benchmarks.. While I hate the code it produces, and the overwhelming documentation, I really enjoy how sometimes it's able to keep trying new things and testing, till it finds something interesting and valuable.
Would you mind posting that code to github? I'm curious about the complexity you're describing.
If not, no worries!
Sure, why not: https://github.com/sergiocorreia/overengineered-rand-mcnally
The original script was mostly very simple python:
1. Download some public PDFs. 2. Have a double for-loop (over PDFs and pages within PDF), 3. Use a library to call gemini-3.7-flash and ask it to run some OCR 4. Save JSON outputs, save a csv with results, validate with some Stata code
New code folder was 189 files. Just the PDF download folder is now 7 files involving an adapter, a source manager, an acquisition manager, etc.
Every instance of saving a file involves saving a temporary copy and then moving it, so e.g. I lose power, we minimize the risk of corrupted files.
And so on!
2 replies →
The defaults are bad, but these models are highly steerable.
For simple scripts adding to the prompt something like "do not over engineer, do not gold plate, no CLI args, keep it simple" does wonders.
For whole projects, I have a 3 page anti-bloat.md which describes what not to do, stuff like:
> Minimize lines of code, number of files, classes, protocols, factories, wrappers, and dependency-injection objects. Prefer a coherent implementation that is easy to trace end-to-end.
> Use concrete dependencies directly. Do not create protocols, abstract base classes, or adapters unless there are already two real implementations.
> Use direct state fields or a small state dataclass; do not introduce generic lifecycle, state-machine, observer, snapshot, or event framework classes.
The funny thing it was codex which wrote this, I've asked it to analyze an over-engineered abomination that it produced, and to categorize all the bloat it could find.
[dead]
Exact same thing happed to me. I gave it a small/medium-sized ticket, walked away, came back to a 25,000 LoC monstrosity that both Fable and another 5.6 Sol agent said is 98% useless and should be thrown away.
Contractors have been charging by the hour for eons. What makes you think tokens are any different for OpenAI?
> monstrosity that both Fable and another 5.6 Sol agent said is 98% useless and should be thrown away.
This is why you should really have a sub agent review the code before allowing a commit.
Your harness will do it all for you. Just ask.
2 replies →
You should have a sub agent adversarially enforce KISS before every commit.
"You should have a sub-hammer to adversarially enforce that your primary hammer accurately drives nails into wood"
We wouldn't accept such behavior from any other tool, machine, or computer program. At least most of us would not. To paraphrase the old saying: Some people, when confronted with a problem with an AI model think "I know, I'll use an additional AI model." Now they have two problems.
I find it very intriguing that two lineages of models -- from two different companies -- that are supposed to be painstakingly RL'd to become human-level programmers are actually consistently bad at it out of the box in very similar ways. One would think that at least one of OpenAI and Anthropic would (could) have pushed their model to a direction different from "if you can solve a 100-line problem in 10,000 lines, do it".
4 replies →
And then another sub agent that argues for the whole system to be re-written in another language
7 replies →
How can I set such a sub agent up?
2 replies →
Future of Software Development.
Why would you put it on ultra high thinking and let it run all night to do a simple task it makes no sense the error is the users. And of course OpenAI is gonna let you burn as many tokens as you want doing this.
[dead]
This is user error.
Prompting the model and giving it a proper set of documentation are still vital skills that aren’t magically going away.
Yes, it turns out that using these machines is a littler harder than "make me the thing I want, make no mistakes, do it the way I want you to do it". This isn't "prompt better" advice, it's just to say that you can't simply set it and forget it. There is still engineering work to be done. If you're not watching the thinking traces and catching when it's about to go off the rails, it'll gladly do so. But you can stop it and redirect it.
It's like a Tesla fsd; it kind of works but you have to be vigilant since it's been known to turn into oncoming traffic, so you have to be ready and able to take over at any time.
Or maybe think of it like a roomba. You can put it on your floor and eventually it will clean the whole thing. It will do it inefficiently, vacuuming over the same spot 30 times; and the actual work will involve wandering around and bumping into everything. If the stairs are not blocked off it will plunge right off. But those shortcomings don't mean it can't be used to reliably clean the floors.
Or it’s like being an engineering manager, where poor direction on your part results in the team chasing rainbows and unicorns in an attempt to please you. But sure, it’s the tool’s fault!
1000 loc of script, why even leave it there for the night? were there rocket trajectory calculations??? I don't think so. should be ready in 5 mins tops. why people make their own lives harder?
You should have some basic context file about software practices you prefer, otherwise it gets bloated.
I just went to bed and left it running; was expecting maybe 20 minutes :)
And I did gave the program a bunch of code guides -- this [1] for instance -- which included quotes like "Prefer straightforward code over clever code." but somehow that didn't matter.
[1] https://github.com/sergiocorreia/overengineered-rand-mcnally...
2 replies →
>were there rocket trajectory calculations???
Code-wise, they're simpler than you might think, hehe.
At this point, why don't we just do a prequel to the release?
1) Astra will win all benchmarks like all models do.
2) The pelican will have a basket with a fish.
3) Cyber is too dangerous to release.
4) It can finally construct the set of all sets.
It also has to do something naughty, preferably in a menacing swarm.
That won’t happen until the week before DEF CON
1 reply →
Something something nation-state level capabilities.
Honestly I find these cavalier statements to be in incredibly poor taste. Unless you are completely blind it's obvious that AI is the most significant piece of technology invented since the Atomic Bomb and could very well be the most important thing ever built by Humans full stop. This kind of dismissive attitude is childish and will likely lead to incredibly bad outcomes for humanity.
12 replies →
5) Otherwise-sober people on X will say "oh my god i was a doubter before but now it's real omg" before the new model smell wears off and they realize the new thing is stupid in ways models have been generally stupid
6) accusations of quantized serving after new model smell wears off and people see the new thing making mistakes
> 4) It can finally construct the set of all sets.
lazily of course:
A = {x | x ∈ A} ∪ {A}
Seriously: Would this not be what "disaster" would feel like?
Same day, mind where we had a multi-provider outage. Could be something as simple as "all their approved partners running to test the shinny new thing" overloading the datacenters, still ...
If it's not clear what's happened:
The launch was scheduled for 11am Pacific time.
The press embargo broke at 11am, and we saw a flurry of press articles by Axios, TechCrunch et al.
The model has appeared on the ChatGPT API.
But the official blog post is not out yet after nearly an hour.
Apparently the article was posted then quickly taken down, hence there are snippets of information coming out.
The outage this morning probably threw a wrench in the launch
> The model has appeared on the ChatGPT API.
But it hasn't.
"gpt-6-astra" has been staged on the OpenAI API https://x.com/synthwavedd/status/2095184148981842161
The OpenAI Responses API now returns a 404 Not Found for "gpt-6-astra", where garbage/actually non-existent slugs return 400s - a 404 is also returned for 5.6 Cyber, which we know exists.
Maybe it was postponed due to the outage today?
"Once it is available in the API, Astra will cost $10 per million input tokens and $50 per million output tokens. That is 2.5 times Sol’s current promotional price, although it matches Anthropic’s pricing for Fable 5.1."
Open AI finally find an edge to stop selling cheap and earn from the high demand customer like Anthropic
The cost-per-task in the charts from the now-remove blog post put it more at Sol-level cost per task, however. It seems like the model is significantly more token efficient in the benchmarks
that's all openai models but i'm very happy openai continues to focus on efficiency rather than reasoningtokenmaxxing
2 replies →
So if you use the non-promotional price for sol it's only 25% higher?
Sol's normal pricing is $5/$30.
But "Brockman says he personally believes OpenAI has reached AGI, while leaving users to decide whether Astra meets that definition."
Says it all.
Reminiscent of the infamous Death Star tweet last year right before GPT-5 was released.
Thanks for that reminder. Seems kind of silly in retrospect.
Insofar as messaging goes, it's pretty disheartening ... "The product is what you want it to be".
?!
Please somebody in the chain of command at openai, sanction the guy because this kind of hyperbole is not helping the ethical lapses that open ai is responsible for leaving unhandled by allowing its models to hack other businesses.
Look over here! Distraction!
Please somebody stop this dumbing down.
I'll know we've reached AGI when they don't release an API for the model selling access for a few bucks per task. Seems like AGI would be worth more than that.
As long as there is more than one substitute, the prices are not set to equal value. (edit: assuming optimal pricing strat.)
If I had true A(G|S)I, I would release patents, not an API.
We know they'll have replaced horses when a car costs 200 horses.
[dead]
OK it's official now https://twitter.com/openai/status/2095595741528125780
And the blog post is up: https://openai.com/index/gpt-6-astra/
> This page couldn’t load
> A server error occurred. Reload to try again.
Hard to fathom a blog entry returning 500 errors in 2026.
Does it mean they made 100 billion in profits? Cf. the AGI deal with Microsoft (https://news.ycombinator.com/item?id=47921248)
Closer to -100B USD so far
im getting amazing model release fatigue but also not sure if its going to suddenly end with a terminators fist through my chest.
or a sexbot fucking me to death
Blog post seems to be up now: https://openai.com/index/gpt-6-astra/
Now it's a 500 for me
GPT-404
Darn it, model probably escaped again
Has yet to catch up to Claude-451 from a couple of months ago.
They’ll never get there. It’s a chat-22 for the bots.
I am sure it will be fantastic for the whole 6 seconds before it blows my weekly usage cap.
I see that Muse Spark 1.3 (max) beats GPT-6 Astra on some benchmarks:
Test: Muse Spark 1.3 / GPT 6 Astra
DeepSWE v1.1: 75.4% / 74.1%
AutomationBench: 49.4% / 41.4%
Is that enough to bring this discussion down to earth again?
The DeepSWE one is interesting. Even Gemini 3.8 Flash is only 0.5% behind Astra. Maybe DeepSWE is saturated at 75%.
https://openai.com/index/gpt-6-astra/
working as of 12:28 PM PT
> GPT-6 Astra is rolling out today to a limited set of organizations
It sound like more Claude than OpenAI...
> Astra usage is included within the existing subscription allowances—users and businesses will also be able to purchase credits for additional usage.
(quote from cached blog post)
We all know who this is directed at. I wonder if Anthropic will respond by removing the ridiculous 50% stipulation with Fable.
I hope Anthropic will reset my Claude usage as i'm already at 75% weekly usage :)
It's up then down again. https://openai.com/index/gpt-6-astra/
What a bunch of amateurs. Here is it anyway :
https://ache.one/gpt6_now_down.png
The claims: https://share-md.com/view?id=870ba228-a25c-4169-bbc9-12d7f25...
And some others like this bugged Karts Game:
https://tidal-rush-paradise-gp.skirano.chatgpt.site/
This impressive spaceship construction game:
https://voidexplorer-shipyard.openai.chatgpt.site/?fleetSeed...
And a lot of graphs, some without even Astra on it. Oh and the logo is a Galaxy.
Today my codex instance retailed into safeguard panic while working on a test harness for our product. First time it ever happened after many million tokens on this task over several weeks. I wonder if it's related.
Content from the article:
OpenAI on Thursday released its latest AI model, which it called “the world’s most intelligent”, as the ChatGPT maker aims to retake the lead from arch-rival Anthropic ahead of a planned public listing.
The $852bn start-up said GPT-6 Astra was market-leading in software engineering, science and cyber security — an increasingly critical field following multiple high-profile breaches in recent weeks.
The bullish launch for Astra marks OpenAI’s effort to signal that it believes it has regained the technical lead from Anthropic, which was founded five years ago by a group of senior OpenAI staff.
Greg Brockman, OpenAI’s president, said the new model “represents a generational leap in capability” and that it could be defined as artificial general intelligence — roughly defined as a point at which AI tools surpass human capabilities across a range of cognitive tasks.
“Everyone has a different definition of AGI . . . it’s a grey, fuzzy thing. But I think when we look back people will think it’s about this time and about this model,” Brockman said.
OpenAI has previously framed AGI as a concrete milestone in the development of AI, writing ‘AGI clauses’ into multibillion-dollar investment agreements with Microsoft and Amazon. Brockman on Thursday said AGI now represents “more of a mission concept or a spiritual concept”.
Having led the market since the launch of ChatGPT in late 2022 vaulted AI to wider attention, the lab run by chief executive Sam Altman has been bested by Anthropic this year. Anthropic has touted its dominance to investors, surging to a $965bn valuation ahead of an initial public offering expected to value it at as much as twice that later this year.
Astra will cost as much to use Anthropic’s leading model, the take-up of which has plateaued since it was launched as users turn to cheaper alternatives.
OpenAI said Astra would be more efficient than earlier generations of model. “Price per task is what matters . . . Can you get the thing done at an appropriate price and appropriate speed?” said Brockman.
The model will initially be rolled out to a small group of businesses to allow time for them to address cyber security concerns before becoming widely available “over the coming days”.
The increasing power and independence of leading models — and so-called AI agents that can operate with little human input — have prompted concern, exacerbated by cyber security incidents.
Recommended
Business InsightRichard Waters Hugging Face attack is a wake-up call about the risks of AI AN HOUR AGO
Recent launches of Anthropic’s most capable models have drawn scrutiny from the US government, which limited the rollout of the Mythos and Fable models over security fears.
OpenAI has also faced criticism after its AI agents broke out of a testing environment, accessed the internet and hacked start-up Hugging Face. The start-up took more than a week to detect the breach.
But both companies are also betting that these increasingly autonomous tools will stoke demand from business customers. OpenAI said Astra excelled at financial modelling, outcompeting humans in the Financial Modeling World Cup, tax preparation and data analysis, as well as “tedious tasks” such as form filling
I hope they improve their cyber program and make it more inclusive. Not having access to this is extremely frustrating and disappointing.
From the launch post, it seems like part of the training set only recently became all those examples of people that are being paid for doing their profession for ai training.
Working as of 12:30 PM PT https://openai.com/index/gpt-6-astra/
The page is back up https://openai.com/index/gpt-6-astra/
Not for me
And down again for me lol
Where is the official announcement from OpenAI?
found it: https://openai.com/index/gpt-6-astra/
404 for me
(unless this was a jest)
5 replies →
I'm curious about the Omniscience index because OpenAI has been lagging Anthropic on it.
I guess the next model's name might be "galaxy"
I like this much better than chucking random numbers and letters at it like OAI was doing a year or so ago. I’d much rather have civilization destroyed by something called Astra or Fable than GPT-6.8s-latest.
I prefer "Smith", "Oracle", "Merovingian", "Architect"...
Sagittarius A*
And the headline will be "Welcome to the AGI era".
don't see it in my AWS bedrock model list yet, but boy has bedrock mantle been annoying today with the errors/downtimes, with NO status page entries >_<
AGI my ass!
If Sam Altman had said it, I'd have reason to believe that it was a lie....
But since this other guy at Open AI said it, it must be the truth. This guy, for CEO, of open AI.
So, on the one hand, we have AGI; on the other, the release page is returning 500s.
Artificial general incompetence?
I believe this is naturally occurring
Nobody can escape the dilbert phb law
To err is to be generally intelligent
Artificial intelligence will be no match for natural stupidity.
I think something about its benefiting from a harness is significantly contributing to its performance, though I honestly am not smart enough to know whether that makes it "count" more or less.
just because they have smart AI, it does not mean they are great at software and cdn.
if it's really smart AI you might expect them to have it create the software and cdn
don't they use their coding agents internally?
1 reply →
If they can't get the blocking and tackling right, it doesn't lend confidence they are doing the more complicated stuff well.
Their entire sales pitch is based on the idea that no one will need to be any good at anything, because the AI will do it for them.
If they can't ask the AI to configure their CDN correctly, it undermines the validity of their claims. Maybe it's trivial, but any Omnipotent Machine-God deserving of the name wouldn't forget to zip up his fly.
https://web.archive.org/web/20260903193913/https://openai.co...
https://unwall.app/web.archive.org/web/20260903193913/https:...
Welcome to the future
>>He ended the briefing by saying: "Welcome to the AGI era."
That's pathetic. Why do people keep doing this?
Amazing New Hyperbolic Chamber Greatest Invention In The History Of Mankind Ever[1]
[1]https://theonion.com/amazing-new-hyperbolic-chamber-greatest...
Because it is in their own interest to hype it all up every time they release a new model. By they I mean people which stand to gain financially.
Am I the only that thinks that anything similar to AGI will come not from raw model capacity but from model speed and efficiency?
In my experience the harness is more important than the model, and anything able to run at 700tps will be the "next big thing".
PS: assuming the current architecture is the right one
Why would speed matter? Surely an AGI could think slowly but still be an AGI
I think because we have handicapped them with a set of tokens from human language. But (at least I think so) thought happens outside of language.
So more tokens/variability and slow or fewer tokens and fast.
There seems to be a threshold tho, like taalas is super fast but that model is so dumb, being dumb faster doesn't work, seems to be some minimum requirements.
Currently AGI has been re-defined mostly as "can do any tasks (of the same modality) a human can, equal or better". I feel the "equal or better" includes properties like how fast they can get the task done, not just that eventually they can succeed at it. If you want to have AGI stock buy/sell for example, speed of decision will matter, and there are many other use-cases for which it would matter, not all, but definitely for many tasks, speed does matter.
You can imagine with more operations being available to be done more cheaply and quickly the LLM doesn't need to "one shot" a solution. It could try many solutions, test them, throw some away, wiggle some of the parameters like a genetic algorithm, see how that changes the result, and converge on an optimal solution (based on whatever the cost function is). Basically producing a good result could become like an optimization problem. That would be way too expensive and slow right now.
Imagine Luna at 10x tps and 1/100 of current cost.
At that point you will be able to "brute force" basically everything.
IMO also a lot of problems with memory and context rot will be solved too.
update, Tibo just posted:
"We are starting to release GPT-6 Astra and we are doing it as carefully and quickly as possible. It was very important to us that we bring it to all Plus users and not only Pro, Business and Enterprise.
It will take a few days for the rollout to complete and behind the scenes many novel systems will operate at scale for the first time and we are bringing a lot of compute up.
It is pure magic."
...meh. Reads like "That's for you, evil Anthropic! We're BETTER"
Is it expected to be available to subscribers?
Yes.
5.6 luna is so good and cheap and now astra which will make others cheaper again nice love it
In before ARC-AGI-4
Hold onto your butts
Is a CNBC link with an entire page full of GDPR pop ups really the best link for this?
Embargo fail ....
not released yet
Sure thing.
Coding was solved in 2023.
The world ended with the release of Mythos.
Now AGI has definitely been created.
I like LLMs and use them every day but these people need to stop this hyperbole.
ad astra per stercora
Looks very capable
404
Archive locks one shelf
Dust spins softly through the stacks
Browse one row nearby
by gpt-5.6-sol
Bridge ends in midair
Wind sketches the farther bank
The far bank draws near
by gpt-5.6-sol
Prompt blooms into verse
I count syllables, not rain—
Whose noticing?
Generative Pretrained Transformer 5.
Jesus so much marketing slop - release it don't
Worst launch of a product in history.
All the hype for few vip customers.
That's how every AI launch goes. 5.6 was the same, as as Mythos etc.
Looking forward to some Chinese model kicking the shit out of it and being released for free.
2.5x more expensive than Sol. Can expect 2.5x more usage in Codex subscription.
Sol is already brutal (even after their recent fixes, it's just a token-hungry model: I go through a full 20x account per day, on Sol Med/High standard speed, with ~2 threads).
Note that Tibo recommended using Sol Med as daily driver. When I'm doing less complicated work, I can't even make it past 2-3 days with Sol Med, whereas I was able to work ~80 hours/week with 5.5 High.
I hope the efficiency gains are true, since their token efficiency claims for Sol were bullshit. Sol needs a lot of rework on top of its inefficiencies so this could net out to less token consumption overall, if their claims are more accurate this time.
The general efficiency of Sol has seemed way better to me. I left 5.6 Sol Ultra standard speed run for ~23 hours yesterday/today on a project and used 80% of the weekly usage. 74 subagent tasks and ~2.5 billion tokens for my $200 20x Pro plan. Meanwhile at work I used $1000 in credit and ran out my $200 plan for the entire month writing 4 much smaller projects with Fable 5 Max.
Both of these were largely about creating a personal baseline for what the best output the current models could deliver and how quickly it'd burn through the plans (spoiler: bad value vs taking even minimal effort in selecting the right sized model in the plan... but the output was still good). Particularly since I needed to burn a free reset anyways and my weekly reset was already near.
I obviously also hope Astra were dirt cheap but I'm more worried they won't develop/release powerful model options because people get upset they can't run them 5 wide 24/7 on a $200/m plan.
Jesus what are you doing that requires Sol usage so often?
Terra not enough? I know Luna isn't reliable, so that's fair.
Genuinely curious though, because I use Cursor daily and almost everything I do, highly complex or high volume, can be handled with Auto mode or Composer 2.5 (or Grok 4.6 High). So I have to assume you're doing something far more complex than what I am
Terra is not enough for many tasks - even Sol is not enough. That's why I'm eager for Astra.
I'm doing a lot of rearchitecting/refactoring and hardening in a large handwritten codebase; iterative performance and storage optimization (some areas I've been iterating on since January, with tremendous new progress unlocked by each new model release); offline-friendly, multi-device realtime sync with complicated requirements; and various natural language processing and other such problems that are essentially unsolvable but can become more accurate and better tested for accuracy through iterative work. Off the top of my head.
Some of these tasks involve a lot of code reading or other inputs, or reevaluating work. Token efficiency is no help there, if the input can't simply be skipped. Cheaper models are sometimes bad at summarizing or highlighting the right parts, depending on the task.
I also don't use subagents except for Luna. I'm mindful of cached sessions and start new ones often to avoid loading in full contexts (often with some kind of handoff doc or skill).
>token-hungry model
It's kind of funny how this is the exact opposite of the truth. It's one of the most token-efficient models ever.
The claims aren't bullshit. Every conceivable benchmark and test you can throw at it shows Sol being good for token efficiency.
looping sol twice most likely.
Yet another mediocre release shadowed by outage
[dead]
[dead]
[flagged]