fable 5.2 > gpt 6 astra tested both models with same prompt at high reasoning and results came out really different fable next model entered stealth testing today, so i don’t think it's coming out anytime soon from early testing: > huge step up from current fable > output... looks better than astra > slower and expensive still running more tests and will drop the outputs soonshow more

here's full comparison of all outputs against fable next model name is not confirmed so i assumed it's gonna be fable 5.2 but let's see

Adding glowy lights doesn't mean is better. Thats a very vibecoder point of view

it needs to look visually good so yeah

The bottom lights litteraly glow in square shape

Astra won’t like this

haha astra will have next model soon too

the fuck is fable 5.2 man hahahaha

try same prompt of tibo on fable 5.1 and if it knows you should have something either new checkpoint or new fable but not everyone have it and started earlier today

these mfers

they will give you few days after everyone enjoyed it lol

whne output from Opus 5.2? Lemme try it with Opus 5.2 Send me prompt G :)

i posted it days ago it's better not same as fable 5.2

hmm That's good then

id assume 5.2 to be a legit step up... damn man any thoughts when? I am getting bored of astra

so i did just a little testing cuz it's expensive opus 5.2> fable 5.2> astra 6 (pretty sure astra next model will out soon if fable 5.2 come out)

fable 5.2 mogs Astra

yeah but for now astra leads

Bro.. Fable is 5.2 looks GOOD

posting frontend test really soon

How are you getting these early access lol share the sauce

i posted it already how check if it answers you question tibo one in claude code or chat or co work then you have it

i can already see this in Fusion, this looks mad good

haha soon soon

Wonder what your prompt was hahahahahha I've been doing basically the same test for a while now. The real magic happened when I asked astra to actually create a landscape out of it.

yooo wtff i want prompt for this

one promt and y'all limits on gpt 6 astra just cooked so sad hahahaha btw, fable 5.2 looks interesting here

if anthropic release new fable openai will release new astra too

@notjazii interesting to see the different results! makes me think about how much the underlying architecture impacts outcomes.

Rest in peace to all those people who keep jumping from Claude to GPT just because their models are good once in a while. Right now there's no point in staying with ChatGPT because their usage limits are worse than Claude's and Claude can come back with a better model within a week or two of an OpenAI model release. At this point I would say just stick with Claude and go through ups and downs rather than going through ups and downs in both Claude and OpenAI

do you think they will come out this week?

opus 5 next model have good chances for today but fable nah at least 10 days i'd say

I hope for a reset, I need it, I have used up almost all my weekly credits in a day, I am working on a new gaming platform.

Woah fable mogged Astra

Quality step is welcome. Still doesn't help if normal users can't afford to run it daily.

This is Anthropic's response to GPT-6. They were waiting for them to release it and now they'll mog

Astra looks better for me

Also those glowly lights don't even look good, lol

One prompt at high reasoning is not enough to call it. If the seed and the reasoning budget were not pinned, most of that gap is sampling noise. Five runs each and compare the spread, not the best output.

Bro is cooking hard, staying tuned for more outputs :)

The 'high reasoning' setting isn't standardized across providers. When I ran cross-model comparisons, I found one model burning 3-4x the tokens to reach its answer. The quality gap looked real until I accounted for compute.

As always fable adds good default taste on its own, for astra you need to give it

btw Jazii wanted to ask, did you manage to qualify for new creator revenue?

Any leaks for Fable 7?

@grok does anyone have access to Fable 5.2 yet?

glow up or glow down, speed matters more

Reasoning level needs stating in any model comparison.

So how is this better than Astra? Because of light? Look well into both, why should there entrance be packed together on fable, see how clean and practical Astra is

Same prompt on both, that is the fair way to compare.

I got the opposite result in my tests. Astra still works better for most of what I do, although Fable looks much better than before

Fable 5. 2 is the banger Model for me. Astra is good but it use tokens too much.

Like these vids. Astra absolutely insane

the lighting makes a massive difference