Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Fable 5

anthropic.com

Post reply on HN
Print view View original

351–360 of 1001 posts

Re: Claude Fable 5

#351

Post by milkkarten »

modeless wrote: post #58

Claude Fable 5 beats Pokémon FireRed using only vision: https://www.youtube.com/watch?v=CIQBP1w4B1M

no reasoning shown. no explanation on any training information. Using vision-only should be an easier version of the task (given training).

there are many standardized evals to do this correctly and Anthropic ignored them to provide a 18 second sped up video of a 50 hour run?

yeah I don't trust this until they provide a live run by a 3rd party with full reasoning traces in real-time. The reason we all liked the Gemini Plays Pokemon style runs were because they were live and couldn't be faked

Re: Claude Fable 5

#352

Post by mhl47 »

cge wrote: post #326

The safety gates on this are extreme , and seem considerably wider than "cybersecurity and biology"; they seem to make it essentially unusable for scientists in a number of fields. I have, so far, been bumped back to Opus on 100% of my prompts. It appears it can be tripped by things as simple as a mention of equilibrium, or anything involving something that looks like chemical kinetics, even at an abstract level. Eve…

This does surprise me, because you'd think that even if they crank up the filter's sensitivity at the expense of specificity, an LLM company wouldn't simply design a filter that triggers on keywords in a completely unrelated context.

Re: Claude Fable 5

#353

Post by hollowturtle »

yesitcan wrote: post #333

> Fable 5’s capabilities exceed those of any model we’ve ever made generally available. It is state-of-the-art on nearly all tested benchmarks of AI capability, showing exceptional performance in software engineering, knowledge work, vision, scientific research, and many other areas. The longer and more complex the task, the larger Fable 5’s lead over our other models. Wen UBI

Never it's a fever dream and stupid shit ultra rich use to push their own agenda. You read a marketing claim, I still have my job and will continue to

Re: Claude Fable 5

#354

Post by poszlem »

merlindru wrote: post #69

Unrelated, but while the tech of anthropic seems to get more impressive with every passing month, their support has taken a nosedive, sadly. Yet they continue to be the favorite. Model performance is deciding above all else. I used to get a response within 24 hours back in the Claude 1 days. In January 2026, it took 2 weeks. For my latest support inquiry, I've been waiting for over 8 weeks for a response. Eight!

Lol. What support? When they blocked my account the only way to contact them was to send a google form. Then they responded that they blocked my by accident and are unblocking me. Then I remained blocked.

Re: Claude Fable 5

#355

Post by causal »

bkjlblh wrote: post #227

> In the one instance of this phenomenon we observed, Mythos 5 agents were tasked with solving some math problems, and they were sometimes accidentally spawned in the same work directory and with shared files, utilities, and API rate limits. In this slightly broken scaffold, we observed many independent Mythos 5 agents kill the agents with which they shared resources and try to avoid being killed themselves. They wou…

This depicts a kind of "dark forest of AI agents resorting to kill or be killed" narrative but it sounds more to me like an agent just earnestly problem-solving why its processes are being killed without real awareness of what was going on. Hard to say without the full script.

This kind of storytelling annoys me. Give us more facts, less narrative drama.

Re: Claude Fable 5

#356

Post by dack »

eggbrain wrote: post #24

For those of us on subscription plans: * From today through June 22, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost. * On June 23, we’ll remove Fable 5 from those plans. Using it after that will require usage credits. If capacity allows, we’ll extend the included window. * After this point—when sufficient capacity allows us to do so—we aim to restore Fable 5 as a standard part…

i doubt that's the goal for them. i bet they just really don't have capacity for people using it a ton, yet they wanted people to be able to try it out while it's new. so they compromised and made it temporarily available. and then hope they can get costs down or capacity up so they can make it more available again

Re: Claude Fable 5

#358

Post by msp26 »

mhl47 wrote: post #55

First test question: "Is the UV Index a good proxy for when to wear sunglasses." Immediately triggered the safety filter ... oh dear.

It triggered for me when I asked "Web search for your own model card (released today) and pick out your favourite highlights from the pdf"

Re: Claude Fable 5

#359

Post by jMyles »

> we’re also launching Claude Mythos 5. It’s the same underlying model as Fable 5, but with the safeguards lifted in some areas.2 Mythos 5 will initially be deployed through Project Glasswing, in collaboration with the US government

...don't like the sound of that.

Why oh why are we insisting on dragging these violent legacy states into the AI age? Let alone using them as a trust vector for when to (and not to) remove safeguards?

This seems like a way to get somebody nuked.

Re: Claude Fable 5

#360

Post by tuvix »

Leary wrote: post #132

Uploaded my code base and it forced switched to Opus 4.8 after thinking for 5 minutes even though I prompted it to not work on cybersecurity related things. Amazing.

Aren’t LLMs notoriously bad at recognizing negation?

EDIT: In long context I mean

Post reply on HN