frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: Why not compare Fable 5 with GPT "Pro"? Why compare with GPT xhigh?

1•behnamoh•1h ago
Every benchmark I see compares the super model Fable 5 with GPT 5.5 xhigh, but that's not a fair comparison because GPT 5.5 Pro has existed for a while and consistently delivers better results than Opus 4.8 Max, so doesn't it make more sense to compare Fable 5 with GPT Pro, instead?

FYI, both of them take a long time to respond, as opposed to GPT xhigh which often returns in a couple of minutes.

Comments

ben_w•46m ago
?

"Pro" is the tier of how much you spend each month to access tokens, while "xhigh" is the setting for how many tokens are spent on any given answer.

(Unless I've missed OpenAI having another ill-advised naming decision to confuse things again, like they did with 4o and o4 models existing at the same time).

behnamoh•32m ago
no, there's a "pro" model you can only access on the web.

Concepts and Methods of the U.S. National Income and Product Accounts (2024)

https://www.bea.gov/resources/methodologies/nipa-handbook
1•gone35•27s ago•0 comments

Gordon Wood's Proust

https://thelampmagazine.com/blog/gordon-woods-proust
1•hhs•1m ago•0 comments

This Jacket Pulls Drinking Water from Thin Air

https://news.utexas.edu/2026/06/11/this-jacket-pulls-drinking-water-from-thin-air/
1•ilreb•1m ago•0 comments

How recent employment changes vary with work experience and exposure to AI

https://digitaleconomy.stanford.edu/project/indicators/canaries-dashboard/
1•jay_kyburz•2m ago•0 comments

I'm a Recovering Impact Investor

https://impactloop.com/article/i-m-a-recovering-impact-investor-and-this-is-what-i-ve-come-to-rea...
1•doener•3m ago•0 comments

Stretchy, soft, and sticky: The next gen of wearable and implantable sensors

https://www.caltech.edu/about/news/stretchy-soft-and-sticky-advancing-the-next-generation-of-wear...
1•hhs•4m ago•0 comments

Anthropic's new Fable model has been jailbroken

https://twitter.com/elder_plinius/status/2064776322979676227
1•Khaine•5m ago•0 comments

Artificial intelligence got better at building itself

https://www.economist.com/science-and-technology/2026/06/07/how-artificial-intelligence-got-bette...
2•andsoitis•5m ago•0 comments

'8647' etched into grass on the National Mall

https://www.cnn.com/2026/06/11/politics/8647-grass-national-mall
1•_spduchamp•6m ago•1 comments

How bigger data diminish the value of research to business

https://www.chicagobooth.edu/review/how-bigger-data-diminish-value-research-business
1•hhs•6m ago•0 comments

Richard Feynman was asked in 1985 if machines would ever think

https://twitter.com/jaynitx/status/2065004600105836908
1•bilsbie•9m ago•0 comments

Berlin, Israeli defence giant IAI sign deal for defence innovation hub

https://www.euractiv.com/news/berlin-israeli-defence-giant-iai-sign-deal-for-aerospace-and-defenc...
3•doener•19m ago•0 comments

Scott Pelley Shows How Legacy Media Got It Wrong – and Bari Weiss Made It Worse

https://theintercept.com/2026/06/11/bari-weiss-scott-pelley-60-minutes-cbs-news/
6•hn_acker•20m ago•1 comments

'News' Site Keeps Hallucinating EFF Staffers

https://www.eff.org/deeplinks/2026/06/news-site-keeps-hallucinating-eff-staffers
2•hn_acker•20m ago•1 comments

A Functional Taxonomy of World Models

https://drfeifei.substack.com/p/a-functional-taxonomy-of-world-models
2•gmays•24m ago•0 comments

Bringing Hilton's loyalty program onto a secure AWS platform

https://www.justaftermidnight247.com/case-study/hilton-premium-club-japan/
1•mooreds•24m ago•0 comments

ICE Officers Break Cameras. Cops Steal Them. Welcome to New Jersey

https://www.techdirt.com/2026/06/11/ice-officers-break-cameras-cops-steal-them-welcome-to-new-jer...
6•hn_acker•24m ago•0 comments

SpaceX officially prices shares at $135 in the largest IPO ever

https://techcrunch.com/2026/06/11/spacex-officially-prices-shares-at-135-in-the-largest-ipo-ever/
7•7777777phil•24m ago•0 comments

Euro-Office 1.0 arrives to FOSS infighting: 'Compatibility is not sovereignty'

https://www.zdnet.com/article/euro-office-is-here-libreoffice-supporters-arent-happy/
5•devonnull•24m ago•0 comments

Silent Android feature scans photos for 'sensitive content' – how to uninstall

https://www.zdnet.com/article/android-safetycore-scans-phone-photos-for-sensitive-content-how-to-...
1•josephcsible•24m ago•0 comments

Show HN: A self-hosted market-research tool with publicly verifiable security

https://atlas.freedomcore.io/
1•MaverickAtlas•27m ago•0 comments

What Would It Look Like If the AI Bubble Popped?

https://www.thebignewsletter.com/p/what-would-it-look-like-if-the-ai
2•chmaynard•29m ago•1 comments

Is Your Writing Yours?

https://personalaisafety.com/p/is-your-writing-yours
1•sofiaqt•31m ago•1 comments

Mu

https://mu.social/
2•doener•31m ago•0 comments

Precision Matters in Block Scales

https://constantinides.net/2026/06/11/precision-matters-in-block-scales/
1•matt_d•32m ago•0 comments

UX Collective: The Prompt is not an Interface

https://uxdesign.cc/the-prompt-is-not-an-interface-41b77277681d
1•valisvalis•32m ago•0 comments

OpenAI Prepping for On-Prem Product?

https://ledger.somantix.ai/posts/open-ai-lays-groundwork-for-on-prem-product/
5•bdroopy•32m ago•0 comments

Data Science Weekly – Issue 655

https://datascienceweekly.substack.com/p/data-science-weekly-issue-655
1•sebg•34m ago•0 comments

Building an AI-Friendly Company, Just in Case AI Takes over the World

https://www.paxerp.com/blog/ai-friendly-company-docs
1•robeym•37m ago•0 comments

Agents' Last Exam

https://arxiv.org/abs/2606.05405
2•matt_d•38m ago•0 comments