"no homerS -- we're allowed to have one"

https://amphetamem.es/meme?id=the-simpsons_06_12_71&text=We%...

Yup, same experience here. I used it for one day, spent the next day fixing its lousy code, then went back to 5.6.

API price cuts were obvious once they made their announcement changing how usage is counted.

These moves all make sense when you take into account the enterprise market.

https://news.ycombinator.com/item?id=49889873

Shots fired, half the price of Opus 5.5.

Yes but it was underwhelming, so they seem to have rushed 6.1 Sol out. Also Opus 5.5 may have spooked them too.

Do you need to? Do you always keep up with all the version bumps on the software you use?

It was so underwhelming that it didn't even make it to chatgpt chat interface

Not an AWS customer, I take it? :-)

They may work for one of the big AI labs.

Have we honestly seen that great a leap in the last 6 months, or just better application of what we had 6 months before that.

We are seeing multiple frontier models dropping on the same day and no one bats an eye, because it's more of the same.

Not that they've hit it but that they are approaching it. The time to panic and steer the narrative is before you hit the iceberg, not after

People have been saying this since GPT-4.

The plateau doesn't have to be perfectly flat, but it's not a straight line upward anymore either (kind of like our work on transistors, where we've kind of hit the bounds of speed in clock cycles but are improving on miniaturization and power efficiency)

Nvidia can start putting weights in silicon if model development slows down.

I think they are hitting compute restrictions. And buying compute right now can be 3-4X. And the costs are increasing. If they train a larger model and demand is high, that’s a lot of compute for Codex subscriptions, which is a loss leader for them. Especially Pro 20X which they just nerfed to 10X.

Another piece of evidence on the pile that the sudden panic and desire to "slow down" is because they're hitting the plateau on capability

I think it's more a token-cost-demand plateau. They've reached the scale and investor trillions to which they can't 10x the hardware cost of inference any more. They can't afford to compete by eating costs and there isn't appetite for more expensive inference.

So in order that they don't bankrupt each other they're looking for the legal cartel behavior coordinating a stop to growth by convincing governments to regulate them into stopping.

There's a lot of juice to squeeze in efficiency but only so much whereas it seemed like capability was going to continue to scale with parameter count.

Maybe it's good news for everyone that model capability is now going to scale on semiconductor cost meaning huge players are going to be very motivated to make semiconductors cheap.

This is literally the plan, open weight models are something like 60% of token spend, and it will get worse. many companies now have model gateways where you can slot in cheaper models via cli for cheaper. we've been using glm 5.x and it's pretty close to SOTA frontier models.

it's also why there have been so many calls for regulation and slowdowns.

No because then I'll go to the competition.

Because it is more work?

Switching models is _very_ expensive in compute (you have to rerun everything from the beginning), and highly variable in cost. Cursor tried doing this for awhile, but inconsistent performance/usage means most users turned it off and pick models specifically.

I guess the question is, does the Dunning Krueger effect apply to models? The dumb ones might think they're up to the task.

These models have a knowledge cutoff that don't just prevent them from knowing about themselves (especially since most data about the model doesn't even exist until after the model is created), but they also don't know about other recent models. Sure, they can search and use other sources, even make some guesses based on the models they do know, but their default stance is more akin to "User asked about model X, model X doesn't exist, maybe it was an hallucination or mistake, let me do a web search...", but that assumes they have web search and are willing to spend tokens on it.

Personally I've taken to having a list of 3 to 4 models in default context with some ordering on which to prefer. Things like GPT 6 Luna is cheap very cheap, use it. Because otherwise the model will assume Haiku or such is the good cheap model to use.

The speed I'm having to update that document has not gone unnoticed.

No.

The old "the bigger number is better", GPT announces model 6.1, the obvious thing to do next is to announce Gemini 27, and after that Claudé 3000, then a flute album.

What's that in summary?

Which paper are you referring to?

No, we're pacing ourselves to have the time to evaluate the impact each new model could have, obviously.

I can't recommend Chinese models enough. My personal favorite is DeepSeek v4.1 Flash but I have tried Qwen 3.8, Kimi 3 and GLM 5.3 which are equally impressive but DeepSeek is the cheapest and fastest regularly hitting 270 token per second.

And yeah I have worked with Anthropic and OpenAI models, they're good but they cost a fortune while Chinese models are already really good at a fraction of the cost.

Probably just the singularity, no big deal

Its a news cycle more than anything, and its ONLY going to get much, much worse. Daily releases, or multiple daily, 30-45, by EOY. Welcome to RSI!

Competition

If you look closely at various benchmarks, you'll see that often models will improve in certain areas while regressing in others. It suggests we're already at the point of diminishing returns.

Probably one of the factors. Signed up to openai pro a few days ago, deciding between openai and anthropic, then sonnet 5.5 was released and am wondering whether I made a mistake.

Luckily it's not a mistake as now we have access to . . . dots.

(and sol 6.1, it seems)

This is it.

It's because they need subscription money and interaction data and so keeping a version bump in the wings to stop the bleeding from your competitor's version bump is the logical thing to do. It has nothing to do with RSI.

Anthropic’s IPO?

New models are distill from the actual unrelease frontier models. They are just giving us better checkpoints.

The initial response to 6 Sol was bad, and Opus 5.5 was definitely winning the public vibes war. Makes sense to rush something out

They're releasing Sol 6.1 because 1. Astra 6.1 got postponed 2. Sol 6 is shitty 3. They have to release _something_ in response to Opus 5.5

Versions is marketing, snapshots/minor variations are easy and the number must go up. Release timing is another OAI's marketing tactic.

RSI

Recursive improvement doesn't imply increased rate, another word for it is "iterative" but this probably sounds too boring for some.

It is the only way to reduce prices while making it look like a good thing.

Aka "we made an oopsie last week and released what should have been called GPT 6 Terra with the name GPT 6 Sol"

It's relentless, isn't it?

GPT‑6.1 Sol matches GPT‑6 Astra at roughly one-fifth of the cost

Astra is a pretty impressive model. Excited to try this.

I love free market competition. We're getting insane advancements every day. I remember when llms used to cost an arm and a leg for decent intelligence

This is great. But maybe part of the motivation is that 6-Sol wasn't as good as initially advertised so they needed to tweak it. I felt a clear degradation in quality in some simple refactoring tasks vs 5.6-Sol.

Guys, are we slowing down yet?

The energy costs of the cloud computing required for gaming are substantially less in power - not to mention overall demand - than LLMs. Come on, we're not in the same energy ballpark here.

Given that the number one cost of inference is memory and compute, and the incremental cost of each is energy, cost per inference is roughly proportional to energy consumption.

Did you care about this when it came to your other computing needs? What PC/laptop are ypu running and how efficient is that?

Look at Neuralwatt. They report energy usage with every call as well as aggregate statistics.

A night of no sleep this one will be.

Okay, now price cut 6 Sol (and rename it to Terra again).

Sorry, I'm GenX. Growing up they showed us "Old Yeller" in the school gym every year like that was some kind of treat.

The misstep was naming it Sol - it was Terra-level all along.

I wonder if releasing this soon sort of validates the rumor that Sol 6 was just the Terra model they bumped up and slashed the price.

Then Opus 5.5 caught them off guard and now they're actually releasing the correct sized model.

You need to use the Blender MCP. There is an official plugin for this now, so the third party one can be avoided.

I've only dabbled but yes with SOTA models it is very good at animating and really most Blender tasks you can think of. Certainly if you are coming at Blender at below expert level it makes it far more accessible and fun to work with.

There are still rough edges of course. But try the official MCP out with Astra and judge for yourself.

Have you tried fable? (I did small experiements and was satisfied, but maybe there are reasons to switch?)

Try it, it's that good compared to openai current offering.

I get better results and usage our of my $20 claude sub than my $100 openai sub... it's that ridiculous

Only takes 5-10 minutes to test your favorite one shot comparison prompt.

Astra 6.1, this is Sol 6.1

That model was implied to be GPT 6.1 Astra, not Sol.

I see where you are coming from. But 6.1 Sol seems like a new frontier in pricing, not intelligence. I do think the deceleration stuff was mostly bluster, but I don't think this release in particular contradicts it too much.

I get decent results telling it to use Luna subagents for implementing commits.

Cache is priced at $0.1/M, 50% as sol 6 and sonnet 5.5.

$2/10 is pretty cheap for a frontier model...

I stopped using LLMs. I shit you not. My life got better.

You live in a ping-pong.

Why they are not even benchmark model against Anthropic or anybody ?

GPT 6.0 Sol was so terrible—I wonder if 6.1 Sol will be good?

Is this due to a similar safety concern or just because it's not ready yet?

The coverage around 6.1 Astra seems deliberately playing into the dubious "safety" narrative in a way that feels distinct. But you may be correct in which case, I would take the correction on board and maybe suggest a different alternative.

Although in theory if OpenAI was boycotted in this way the market pressure would force them to release. Then everyone moves over there. Then Claude faces the same pressure. So even so, I think it could still work even if you have to trade off who you are boycotting from time to time.

Without more details on the credibility of the safety concern this seems like a totally coherent action for customers to take. We shouldn't put up with teasing.

I am afraid to ask for the price multiplier here. And it will burn your weekly allowance not in 1 day, but in 3 hours now? Or just one?

i'm watching the keynote

Still not available to me

Are you guys in dev day? did they start?

Other comments have already addressed this.

the race to the bottom on models is well underway. huge IPO's only really make sense for DC/HW lockups, and going vertical.

It is hard to trust these scores. GPP 6 Sol has been so bad for few days.

They need to fix Astra first. My main issue is with GPT in general is that unless steered it goes into building AI “sloppiness”/machinery that is not “needed”.

The good part is that this kind of behaviour also makes it good to find subtle bugs or debug issues that Fable/Claude just cannot get/fix even when you point it.

How does this jive with the exponential growth claims? Theoretically sol models are better than the 4 series models I was using at the beginning of the year, but in practice the results don’t seem to be much better. They always nerf the models over the course of the release so it _looks_ like the next version is better but I haven’t seen actual capability growth since ~January, and I’m pretty sure that was all tooling/harness improvements.

DeepSeek and GLM made it impossible for "sota" to price any way they want.

GPT 6 Sol is obsolete after only one week! I am glad that they are not afraid to update the model more frequently. The Navier-Stokes thing revealed that it took them only a week or two to train a model more capable than Astra, and I want the pace of public releases to keep up with that.

I got a popup in my Codex just now saying "Try out 6.1 Sol!" and so I clicked the button to try it, and intriguingly, it set my model selector to "GPT-6 Astra Light" which makes me think 6.1 Sol may be in some way just a lighter/distilled version of Astra? defo interesting, not sure if I should read too much into it though. I see now option for directly selecting 6.1 Sol in my Codex Desktop UI.

Came here only to check if the pelican spam has made it to the top again.

“OpenAI's new Pro 500 plan offers OpenAI's highest usage allowance and comes with access to its new "Ultrafast" feature — it also costs $500 per month.

At the same time, OpenAI is also making its existing $200 Pro plan less appealing. In Codex and Work, $200 Pro subscribers will see their included usage decrease from 20x of what the company offers to Plus users, down to 10x of that same allowance. In ChatGPT, meanwhile, GPT-6 Pro message caps will decrease from 200 to 100 per week.”

https://www.engadget.com/2272106/openai-adds-dollar500-pro-s...

Yikes

They released GPT 6 Sol literally 6 days ago. We've accelerated to a weekly model release cadence. That seems like...a big deal.

Typically how long does codex take to update with the right model metadata for the release of a new model?

{"type":"item.completed","item":{"id":"item_0","type":"error","message":"Model metadata for `gpt-6.1-sol` not found. Defaulting to fallback metadata; this can degrade performance and cause issues."}}

OAI and Anthropic are locked in an unwindable race. Fighting over the same pool of users while cost is growing at geometric pace.

Impressive improvements, but GPT 6 Sol came out 7 days ago, and this one will behave differently. The panicked pace is becoming a liability, maybe they should have waited and released this as the 6.0 release

I guess they released this because GPT-6 Sol was underwhelming, they didn't even release it to ChatGPT. It was basically GPT-5.6 Terra for the price of Sol. However, who doesn't like price cuts? Astra for the fifth of the price? Wow, OpenAI have been quite generous recently, I still have not forgotten their 90% price cut with GPT-5.6 Luna, and now this? Astra was truly a milestone, and now they are offering similar "intelligence" for cheaper price. Incredible.

One thing I wish was better communicated is the mileage we get for our subscriptions. I do not fully understand how much usage I get with each model and their reasoning effort on 5h and weekly limit in Codex. I am asking because I know switching to Astra would consume my 5h usage limit quite rapidly, so I avoid it. If I knew how much mileage I would get from each model and respective reasoning effort, then I would be able to plan my workflow better and know when to upgrade model for a task. In almost all cases, GPT-6 Luna (XHigh) have been enough. That's why I appreciate its discount, because its dirt cheap, yet highly capable.

In other news:

In the coming days, we’ll also offer GPT‑6.1 Sol Ultrafast , with up to 8x faster token generation compared to its standard speed in Codex.