To clarify, it is not two models. Its a single model (Gestalt). The other bar on the benchmark is a human created dataset. Suffice to say we didn't expect anyone to beat it, but here we are. And no, it is not Contra model.
I don't understand, where are those models available in stealth mode? Usually people do this on openrouter but I couldn't find any model called "gestalt". Does anyone know or is this just black hat marketing for this contralabs platform, and the model does not even exist?
The name is probably a clue. I wonder if someone just bolted Fable to Astra in a loop? I do that when I'm stuck, and get great results that exceed what either can do alone... Though it is wickedly expensive.
We'll eventually find out what insane heroics they pulled to perform these propaganda stunts.
I'm reminded of stories about how the Russians stole the US Space Shuttle designs and tried to pass it off as a product of Russian and Communist superiority to Western democracy and free enterprise.
It turned out that Buran had unlimited funding status, that villages went without water for weeks before static fires, and that they faked the auto-land system by recording dozens of fighter pilots performing landings and then having a human choose one to play back on the day.
I'm seeing echoes of that in these endless "feats of strength" we're seeing in 2026.
> It turned out that Buran had unlimited funding status, that villages went without water for weeks before static fires, and that they faked the auto-land system by recording dozens of fighter pilots performing landings and then having a human choose one to play back on the day.
Wait: I agree with you that soviet Russia was full of shit and didn't have 1/10th of what they did pretend but...
Are you saying modern day China doesn't, for example, have entirely automated factories building millions of batteries of year, in the dark (for there are no humans, they don't even need light)?
Or that Chinese EV cars flooding the EU market is fake news?
Like: do you literally believe China is behind technology-wise? That China didn't launch 20x more rockets in space than the EU?
Stuff like that? All fake? Their modern cities are all fake news too? They don't exist?
Or you do you believe China has all that but, somehow, it's impossible for them to compete with US AI labs?
It is interesting that the Kimi model gets a very high “slop score” but “professional critiques” gave it a pretty high score. Do critiques have no taste or the “slop” is not slop anymore?
If slop score is mostly “looks like AI”, Kimi scoring high there while human designers rate it well isn’t that weird. Generic can still nail the brief.
To clarify, it is not two models. Its a single model (Gestalt). The other bar on the benchmark is a human created dataset. Suffice to say we didn't expect anyone to beat it, but here we are. And no, it is not Contra model.
I don't understand, where are those models available in stealth mode? Usually people do this on openrouter but I couldn't find any model called "gestalt". Does anyone know or is this just black hat marketing for this contralabs platform, and the model does not even exist?
Very curious to find out which model that is :O
The name is probably a clue. I wonder if someone just bolted Fable to Astra in a loop? I do that when I'm stuck, and get great results that exceed what either can do alone... Though it is wickedly expensive.
My money is on GLM 6
This Thanksgiving period will be a cambrian explosion of capable models.
Deepseek V5 trained on Huawei Ascend?
We'll eventually find out what insane heroics they pulled to perform these propaganda stunts.
I'm reminded of stories about how the Russians stole the US Space Shuttle designs and tried to pass it off as a product of Russian and Communist superiority to Western democracy and free enterprise.
It turned out that Buran had unlimited funding status, that villages went without water for weeks before static fires, and that they faked the auto-land system by recording dozens of fighter pilots performing landings and then having a human choose one to play back on the day.
I'm seeing echoes of that in these endless "feats of strength" we're seeing in 2026.
> It turned out that Buran had unlimited funding status, that villages went without water for weeks before static fires, and that they faked the auto-land system by recording dozens of fighter pilots performing landings and then having a human choose one to play back on the day.
None of this is true.
Fairy tales Americans tell their kids so that they can sleep better at night. The parents, I mean.
No no, he wrote it so it must be true. You're just still affected by USSR propaganda.
A quick google search when highlighting 'Buran' through 'for weeks' shows this is in fact true, and the below commenters are full of shit.
Wait: I agree with you that soviet Russia was full of shit and didn't have 1/10th of what they did pretend but...
Are you saying modern day China doesn't, for example, have entirely automated factories building millions of batteries of year, in the dark (for there are no humans, they don't even need light)?
Or that Chinese EV cars flooding the EU market is fake news?
Like: do you literally believe China is behind technology-wise? That China didn't launch 20x more rockets in space than the EU?
Stuff like that? All fake? Their modern cities are all fake news too? They don't exist?
Or you do you believe China has all that but, somehow, it's impossible for them to compete with US AI labs?
This is a clever bit of content marketing
Glm 5.5 I bet, but still too early to tell
That web site is insufferable.
Thanks for confirming the site is insufferable.
I closed it before it could finish loading because it was so slow. Which does speak to its insufferable nature.
who is --John Galt-- Gestalt?
Why would this not be GPT 6.2 Astra?
"Random AI lab makes up a new benchmark that puts at least one of its models ahead of Fable 5.2"
Yeah this is 100% an ad
Saying this when none of the gallery examples show the new model's output is the ultimate "trust me, bro" move.
Perhaps soon we will know. I think on of the reasons for such stealth benchmarks is testing the model before uncovering details
the benefit of the doubt some of you give to everyone else but just swallow anything anthropic/oai has to say is insufferable
"on the leaderboard [that we created so we could lead it]"
It is interesting that the Kimi model gets a very high “slop score” but “professional critiques” gave it a pretty high score. Do critiques have no taste or the “slop” is not slop anymore?
If slop score is mostly “looks like AI”, Kimi scoring high there while human designers rate it well isn’t that weird. Generic can still nail the brief.
Producing a work that looks and feels great but with a mess under the covers is a thing.
some obscure design benchmark