Загрузка видео...

Не удалось загрузить видео

На главную

Fable 5 on Hyperagent is producing the most creative, ambitious work we've ever seen from our agents. They're self-improving for hours towards open-ended goals. Visual reasoning has spiked noticeably. Outputs are consistently higher quality than Opus, occasionally at lower cost. 5 of our test cases below vs. Opus 4.8...

1,341,439 просмотров • 5 месяцев назад •via X (Twitter)

Комментарии: 38

Фото профиля Hyperagent
Hyperagent5 месяцев назад

1. Visualize all asteroids in the solar system from NASA data Hyperagent ran Fable 5 to pull data and design the visual. Fable’s version included visual upgrades such as orbit traces and hover details while maintaining performance. See it:

Фото профиля Hyperagent
Hyperagent5 месяцев назад

2. Design a site plan for a 100 acre fitness retreat Hyperagent orchestrated image gen with GPT-Image-2 and Nano Banana to bring the property to life. Fable’s version considered flow patterns, while Opus 4.8 essentially dropped boxes onto a plot. See it:

Фото профиля Hyperagent
Hyperagent5 месяцев назад

3. Reconstruct Apollo control panels from technical PDFs Hyperagent reasoned over the diagrams and documentation to create a faithful 3D simulation. Opus 4.8’s attempt skimmed over key details.

Фото профиля Hyperagent
Hyperagent5 месяцев назад

4. Simulate the supply chain for World Cup jersey sales based on match outcomes Fable took a global view. Opus 4.8 narrowed its scope to US only and shipped glaring visual bugs (flickering).

Фото профиля Hyperagent
Hyperagent5 месяцев назад

5. Show the effects of solar flares on aurora Hyperagent used Fable 5 to research and design a full simulation. Opus's artifact didn't load, despite claiming to have tested the work.

Фото профиля Ziwen
Ziwen5 месяцев назад

Looks on fire!!! I also tried the Minecraft prompts and one shot it

Фото профиля Lydia Hallie ✨
Lydia Hallie ✨5 месяцев назад

This looks SO good 🔥🔥

Фото профиля Peter Mick
Peter Mick5 месяцев назад

this reconstruction really is mind blowing

Фото профиля Thomas Ahle
Thomas Ahle5 месяцев назад

I have an app built for Opus 4.8 over two weeks. It has a lot of visual glitches. Should I start over with Fable, or ask Fable to fix it?

Фото профиля Beginnersblog
Beginnersblog5 месяцев назад

The Apollo control panel one stopped me. Reconstructing real hardware from technical PDFs is the kind of task you'd normally hand a team for weeks. The fact that it's chewing on open-ended goals for hours and getting better is the part that's hard to wrap my head around.

Фото профиля Yassine Zaanouni
Yassine Zaanouni5 месяцев назад

I can't believe I was able to create this with Fable

Фото профиля Wesley
Wesley5 месяцев назад

🔥 Really impressive stuff! Fable 5 has absolutely cooked Opus on these. That Apollo panel and solar flare viz are something else.

Фото профиля Awa K. Penn
Awa K. Penn5 месяцев назад

Fable 5 looks so realistic. Just wow

Фото профиля Gina Acosta
Gina Acosta5 месяцев назад

This feels like the right direction for agents. Less hype, more work that actually holds up

Фото профиля Voight
Voight5 месяцев назад

Fable, the most desired model.

Фото профиля Galaxyhub Labs
Galaxyhub Labs5 месяцев назад

It's always important to look at the scope of all models, and Fable is undoubtedly leading the way at the moment.

Фото профиля Anissa
Anissa5 месяцев назад

Fable 5 sounds seriously impressive

Фото профиля ᴍᴜʀᴘʜʏ
ᴍᴜʀᴘʜʏ5 месяцев назад

Fable 5 sounds quite interesting

Фото профиля MisterCarrillo
MisterCarrillo5 месяцев назад

Everyone is comparing models. The real comparison is: What can an agent accomplish after 3 hours of autonomous work? That's where things get interesting.

Фото профиля Shahid Wani
Shahid Wani5 месяцев назад

Woww fable 5 sound 😱

Фото профиля Aaliya
Aaliya5 месяцев назад

The use cases are creative and interesting.

Фото профиля Carl downey
Carl downey5 месяцев назад

good

Фото профиля Ryker Stone
Ryker Stone5 месяцев назад

"This is bigger than most people realize. The real story isn't the ban — it's that one government just showed every other government exactly how to do it."

Фото профиля Shiro
Shiro5 месяцев назад

İ think claude is becoming a overrated

Фото профиля The Operator | AI Ops for Creators
The Operator | AI Ops for Creators5 месяцев назад

All the Fable demos are all fancy UI / 3D generations. Is this a 3D Artist model or an LLM

Фото профиля K!A🚀
K!A🚀5 месяцев назад

Fable feels less like a model, more like a working session.

Фото профиля Snow
Snow5 месяцев назад

this is superb

Фото профиля mirajul
mirajul5 месяцев назад

🤖 AI & ChatGPT Guides

Фото профиля TOny-fishkiller
TOny-fishkiller5 месяцев назад

that's amazing!

Фото профиля Michael Suede
Michael Suede5 месяцев назад

wow nice

Фото профиля ZLH
ZLH5 месяцев назад

Amazing

Фото профиля Aditya Rana
Aditya Rana5 месяцев назад

Looks super cool , there is no way it created it in those single line prompts through .

Фото профиля Ori Torii
Ori Torii5 месяцев назад

Amazing! How about building websites? Any examples?

Фото профиля SparkTwo⚡️ はると⚡️17歳のAI自動化屋
SparkTwo⚡️ はると⚡️17歳のAI自動化屋5 месяцев назад

Claude APIで自動化パイプライン組んでいて、Fable 5に切り替えてから複数ステップの処理精度が体感30~40%上がった。Opus比でコスト下がって品質上がるのは設計の選択肢が広がる

Фото профиля Al-Shamus
Al-Shamus5 месяцев назад

This looks impressive

Фото профиля Teknikforce
Teknikforce3 месяцев назад

The biggest win isn't just having a better AI assistant. It's reducing the friction between thinking and doing. The same applies to content creation. Instead of spending hours editing product videos, tools like Vidicom can turn a product page into ready to publish marketing videos for YouTube, TikTok, Instagram, and more.

Фото профиля Shubham Sharma | AI & Tech
Shubham Sharma | AI & Tech5 месяцев назад

All the Fable demos are all fancy UI / 3D generations. Is this a 3D Artist model or an LLM

Фото профиля Faizan Shah
Faizan Shah5 месяцев назад

AI tools hallucinate. Confidently. Every day. Verity automatically fact-checks AI responses with cited web sources. Free Chrome extension 👇 #AI #ChatGPT #Claude #ChromeExtension

Похожие видео

fable 5.1 vs fable 5 vs opus 5 – three lord of the rings landmarks, built in 3d from one image the setup: one reference image per scene, one html file per build, everything procedural – no meshes, no textures, no image files, nothing past Three.js from a cdn. each model reads the picture, writes its own prompt from it, then builds to that prompt in the same turn. three named camera shots per scene on keys 1/2/3, so it can be screen-recorded. run through OpenRouter tasks: 1. bag end – hobbiton from two frames, outside and in. the round green door has to open onto the room you are standing in 2. barad-dûr – the tower and orodruin from one film still. the eye has to move and track the camera, the volcano erupts on a cycle, the clouds never stop 3. rivendell – jerry vanderstelt's painting. sun shafts that shimmer, water that falls without a break, trees that sway on a gust models: Anthropic fable 5.1, fable 5, opus 5 total cost, three builds #1 fable 5 – $14.97 #2 opus 5 – $18.53 #3 fable 5.1 – $22.38 wall clock, three builds #1 fable 5 – 38m #2 fable 5.1 – 92m #3 opus 5 – 122m output tokens #1 fable 5 – 298,592 #2 fable 5.1 – 439,435 #3 opus 5 – 724,418 lines of code shipped #1 fable 5 – 2,885 #2 fable 5.1 – 4,021 #3 opus 5 – 5,161 biggest single build, lines #1 opus 5, bag end – 2,410 #2 fable 5.1, barad-dûr – 1,375 #3 fable 5, bag end – 1,319 observations: • fable 5.1 is the only model that furnished the bag end interior – a live fire, panelling, books on the floor, leaded diamond windows, against fable 5's flat color and opus's dark tunnel. the round door outside opens onto that room, the hard part of the brief • what it costs is thinking room. the 128k output ceiling is a thinking budget in disguise: fable 5.1 burned 102,116 of it on reasoning and hit the wall mid-file. opus spent 109,241 and hit the same wall. fable 5 spent 61,240 and finished bag end in one call – the only one that did • fable 5.1's first pass is not the finished thing. its barad-dûr came back with three defects you only catch by looking at it – nothing a read of the code would have flagged • it is the best of the three at being corrected. handed a plain list of what was wrong, it returned 32 targeted patches over two rounds, every one applied first try, and it worked out one of the causes itself instead of guessing at constants conclusion: nine scenes, 12,067 lines and 1.46m output tokens for $55.88 all in – and the cheapest model was also the fastest, by 3.2x! follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

18,509 просмотров • 2 месяцев назад