正在加载视频...

视频加载失败

🚨 GPT 5.6 Pro first output on the same prompt we are getting started > frontend/ webdev is not solved or improved yet > but understanding increased a lot > it started to take 20-40 mins again like it used to do before 5.5 pro

451,078 次观看 • 3 个月前 •via X (Twitter)

35 条评论

Rijn 的头像
Rijn3 个月前

Damn, it does not look like they solved UI at all. Hopefully this is fake

Zorai 的头像
Zorai3 个月前

Even Glm 5.2 looks way better… really disappointed

Chetaslua 的头像
Chetaslua3 个月前

Thanks @mirochill for running prompts and join our server we are cooking there

ρ:ɡeon 的头像
ρ:ɡeon3 个月前

who tf writes all these community notes??

🥔🥔🥔 的头像
🥔🥔🥔3 个月前

disgusting

Shikhar 的头像
Shikhar3 个月前

Frontend isn't solved

Tim B. 的头像
Tim B.3 个月前

Where is 5.6 available

Cory Schulz 的头像
Cory Schulz3 个月前

seems like the anthropic drama had delayed 5.6. i’m not even sure they can release it now. I think it would need to go through that 90 day approval first.

🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱 的头像
🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱🇵🇱3 个月前

holy shit thats garbage

Hussain Hashim | Building SundayBack 的头像
Hussain Hashim | Building SundayBack3 个月前

@chetaslua sounds like progress! curious if you've noticed any specific parts of the frontend that are still super annoying or time-consuming?

Kushal 的头像
Kushal3 个月前

maybe they can distill glm 5.2

Yashas 的头像
Yashas3 个月前

Oh this means they focused on engineering

Lex 的头像
Lex3 个月前

I pitted Deepseek v4 flash against the glm 5.2 model for this 3D clock coding task, and strangely enough, Deepseek performed better. I used a simple prompt, and I didn't even write it in English; I wrote it in my own native language.

Choblin 的头像
Choblin3 个月前

svg of ancient clock in 17th century From GPT-5.6 Pro

Adel Bucetta 的头像
Adel Bucetta3 个月前

we've seen similar backsliding with other large models too, it's as if they need to relearn the basics every few iterations

Choblin 的头像
Choblin3 个月前

A Pelican Riding a Bicycle Output from GPT-5.6 Pro SVGs are not so Good from this model.

Bryan Patricca 的头像
Bryan Patricca3 个月前

This would be a good one to compare to fable

vibebuilder 的头像
vibebuilder3 个月前

It’s always been very strange to me that the models can be incredibly intelligent at back-end, but piss poor on front-end. Like, yes taste is involved, but there’s a lot of logic involved with front-end too.

Nish 的头像
Nish3 个月前

damn gpt 5.6 ugly at UI

brew23 的头像
brew233 个月前

so not even close to fable

kris.gmg 的头像
kris.gmg3 个月前

looks that fable >>>

Prompt Heat 的头像
Prompt Heat3 个月前

It looks nice. I didn't like the GLM-5.2 design because it was too bright and unrealistic, but this one is cleaner and works more smoothly.

Curline Zephirin 的头像
Curline Zephirin3 个月前

> it started to take 20-40 mins again like it used to do before 5.5 pro @R2Cdev_ Could it be that the one I used that day was GPT 5.6 Pro?

thatguy 的头像
thatguy3 个月前

okay sorry to be one of the people on anthropics side, but 20-40 min for a better result in the low percentages that you can get the same or similar with high fable 5 for a higher price, nothing to use on codex, and still bad ui??? im sorry but fable still my favorite .

Ferbin 的头像
Ferbin3 个月前

sure, understanding improved. but for frontend coding, you traded it for latency. 20-40 min wait and you've forgotten what you were building. was the understanding worth the wait?

omer khan 的头像
omer khan3 个月前

I hope its true

Razex 的头像
Razex3 个月前

Garbage for real bro taking hours for serious project task to be done

J A Z I I 的头像
J A Z I I3 个月前

Wow this is clean bro

Choblin 的头像
Choblin3 个月前

Why is this case? Why is it now taking same time as previous models? It shouldn't.

InfiniteHexx 的头像
InfiniteHexx3 个月前

If it takes significantly longer, that's proof that 5.6 is a new pre-train

Bughunter Geek 的头像
Bughunter Geek3 个月前

Oh, THAT'S why they said they would solve UI only by the END of the year... 💀

GaleForceAI 的头像
GaleForceAI3 个月前

Yeah, that’s my read too. Better understanding, but frontend still not there. And 20-40 mins is a proper tax.

Tornado guy 的头像
Tornado guy3 个月前

Glm is way better OpenAI should not release this dumb model

Moon 🎑 的头像
Moon 🎑3 个月前

But 5.6 models isn't released yet!?

DrstaOne 的头像
DrstaOne3 个月前

looks same but GPT 5.6 one is better

相关视频

GPT-5.6 vs GPT-5.5 on my custom spaceship prompt. I gave both models the exact same custom prompt. This is also the same prompt I previously gave to Fable 5. For context, GPT-5.6 Pro worked for 87 minutes, while GPT-5.5 Extra High worked for 34 minutes and 42 seconds. As I’ve said before, based on great authority GPT-5.6 will be an incremental/soldi improvement over GPT-5.5, not a “Fable killer.” My rough expectation has been that it would trade blows with Fable 5 on some benchmarks, maybe win around half depending on the category, but not clearly surpass it overall. And again fable five will have bigger model smell, but this was expected. After testing this coding output, that view feels pretty accurate. GPT-5.6 is clearly better than GPT-5.5 in several visual areas. The lighting, shading, chairs, object details, and exterior of the spaceship looked noticeably stronger. The scene was also easier to test. I do want to give GPT-5.5 credit though. It built out the rooms much much better and the planets looked better than GPT-5.6’s. It was also interesting that both GPT-5.5 and GPT-5.6 produced better-looking planets than Fable 5 in this specific test. The downside with GPT-5.5 was stability. The game was much glitchier and harder to test compared to GPT-5.6. But when it comes to the core of the demo, which is the spaceship itself, Fable 5 still beat both models pretty comfortably. GPT-5.6 is impressive, but from this test, it looks exactly like what I expected which was a meaningful incremental improvement over GPT-5.5, at least for indie game demos, but not something that replaces Fable 5. In collaboration with Chetaslua

Chris

250,919 次观看 • 3 个月前