Claude Fable 5 Full 319-Page Breakdown — 20 Highlights

• Channel: AI Explained
• Video: Watch on YouTube
• Tags: claude fable5 full 319 page breakdown ai explained

Overview

AI Explained reads Anthropic’s 319-page Fable 5 system card and brings 20 highlights most people missed, covering architecture, safety, capabilities, and limitations.

Key Takeaways

• Anthropic is definitively riding the exponential. at least when it comes to the length of their release notes, they’re trying to kill me man, 319 pages. but more seriously, of course, claude fable 5 is both quantitatively and qualitatively a
• I’ve probably tested the model in about 100 different ways and scoured the results across both famous benchmarks, like the ones that anthropic want you to look at, as well as quiet independent ones and of course my own private
• That’s definitely worth saying. anyway, let’s begin with those blocks, because it might be one of the first things you notice when using fable 5. at least up until june 22nd, when it’s apparently being taken off subscriptions. doesn’t matter
• The release of fable 5, and i had just been having a conversation with opus 4.8 about getting more fermented food for my gut bacteria. i didn’t realize how useful it would be to have sourcrow or kimchi, for example.
• Think that’s heavy-handed, just wait until you dive into the system card, things get really quite wild, and i don’t use superlatives like that often. before we leave the practicals, we might one day get fable 5 back on subscriptions
• Months, reading between the lines of the system card and various interviews, it looks like fable or mithos finished training around february. and of course, four months is a long time in ai, so it could well be the anthropic
• Safeguards. note though, as i predicted in my previous video, mithos 5 or fable 5 is an improvement over mithos preview. safe cards aside, the fable 5 model we’re getting now is an improvement over the mithos preview model that
• Preview. that doesn’t mean it is a significant improvement over opus 4.8. and yes, over gpt 5.5 and gemini 3.1 pro as well. if you can look past the safeguards, it’s clearly the best model out there, though there are
• You to have a visual sense of the power of this model. i asked in a single prompt, write a pokemon clone, but set in the redwall universe, and look at what it came up with. it’s got a soundtrack
• To write, but the game has up to an hour of playtime, i would say. i’ve published it online if you want to play it along too. the model seems to enjoy hard work and more on that later. look
• Hours and hours of agentic research that the model would do, albeit you can never trust it perfectly, but we could just spend the whole video going through such visceral and impressive examples. but that would be too much fun
• Openai or deepseek and want to use fable 5 for frontier llm development, maybe building pre-training pipelines for example. well, anthropic have instituted invisible safeguards, things like steering vectors, prompt modification that will silently steer the model away from effective
• An alias said this is effectively a stun lock on anthropic’s adversaries. it’s some real end game stuff. the reference is to preserve their lead anthropic is preventing open ai from gaining such capability, which brings us nicely to perhaps
• Wish to advance the rate of ai capabilities. if you’ve been watching the channel, i’ve quoted that many times, it was a 2023 quote. well, here they say that yes, we are concerned about the risks of accelerating the overall
• The rate of ai capabilities, 2023, to not wanting to advance the rate of other people’s ai capabilities, 2022. they kind of defend themselves and say, well, we laid out that shift in this february 2023-6 risk report. but if

Transcript

anthropic is definitively riding the exponential. At least when it comes to the length of their release notes, they’re trying to kill me man, 319 pages. But more seriously, of course, Claude Fable 5 is both quantitatively and qualitatively a significant step forward in AI capabilities. Yes, their system card is long, but 9 hours of reading later, let me bring you the 20 highlights also that you may have missed while everyone was having a meltdown on social media. I’ve probably tested the model in about 100 different ways and scoured the results across both famous benchmarks, like the ones that anthropic want you to look at, as well as quiet independent ones and of course my own private benchmark. Before we begin then, what is the TLDR? Well, yeah, it’s a good model. Not if you get blocked, of course, but damn, yeah, it’s good. I haven’t felt somewhat unnerved by a model releasing quite a while, so that’s definitely worth saying. Anyway, let’s begin with those blocks, because it might be one of the first things you notice when using Fable 5. At least up until June 22nd, when it’s apparently being taken off subscriptions. Doesn’t matter if you’re pro or max, you won’t be able to use it. They’re tired of subsidizing poorer users, they want us all to spend usage credits. Pay for the actual cost of this massive model. It was a minute after the release of Fable 5, and I had just been having a conversation with Opus 4.8 about getting more fermented food for my gut bacteria. I didn’t realize how useful it would be to have sourcrow or kimchi, for example. And I asked Fable 5 after switching to the model, review this chat for additional recommendations. Yes, I know I misspoke review. This got flanked as a biology request, so the chat got paused. And by the way, if you think that’s heavy-handed, just wait until you dive into the system card, things get really quite wild, and I don’t use superlatives like that often. Before we leave the practicals, we might one day get Fable 5 back on subscriptions like pro or max or team, when sufficient compute capacity allows anthropic to do so. Second, even if you’re not done philosophically digesting the import of Clawed Fable 5, rest assured that more capable models are arriving in the coming months, reading between the lines of the system card and various interviews, it looks like Fable or Mithos finished training around February. And of course, four months is a long time in AI, so it could well be the anthropic researchers are right now using the next gen model. In case by the way, you’re a little bit confused by the names Mithos 5 and Fable 5 are the same underlying model weights. It’s just that Fable 5 has more safeguards. Note though, as I predicted in my previous video, Mithos 5 or Fable 5 is an improvement over Mithos preview. Safe cards aside, the Fable 5 model we’re getting now is an improvement over the Mithos preview model that sent many into panic back in April, albeit generally speaking, as mentioned on page 50, the improvement is moderate. But don’t be confused, just because Mithos 5 or Fable 5 with the safeguards is only a moderate improvement over Mithos preview. That doesn’t mean it is a significant improvement over Opus 4.8. And yes, over GPT 5.5 and Gemini 3.1 pro as well. If you can look past the safeguards, it’s clearly the best model out there, though there are some nuances in some areas as I’ll get to. Trust me, if I try to tell the art this video, the TLDR would also last about five minutes. Just before we dive deeper into the detail though, I just want you to have a visual sense of the power of this model. I asked in a single prompt, write a Pokemon clone, but set in the Redwall universe, and look at what it came up with. It’s got a soundtrack that I’m using, but dozens of playable levels and interactions you can have with the characters. Of course, it’s also got menus and playable characters and companions that you can bring in to the adventure. The prompt took two minutes to write, but the game has up to an hour of playtime, I would say. I’ve published it online if you want to play it along too. The model seems to enjoy hard work and more on that later. Look at this impressive idea from Ethan Mollick of an isoconic passage shot. Essentially, you can click anywhere on a world map and see how long it would take to get there realistically based on real data from New York City. hours and hours of agentic research that the model would do, albeit you can never trust it perfectly, but we could just spend the whole video going through such visceral and impressive examples. But that would be too much fun because this 319 page system card contains dozens of bombshells, like this one. So you know how they block Claude for requests relating to biology, what about if you’re using it for machine learning research? Maybe you’re a competitor like OpenAI or Deepseek and want to use Fable 5 for Frontier LLM development, maybe building pre-training pipelines for example. Well, anthropic have instituted invisible safeguards, things like steering vectors, prompt modification that will silently steer the model away from effective answers. You could say sabotage those attempts. Again, these safeguards will not be visible to the user. Of course, most of you will not be using Fable 5 to accelerate machine learning, but still, one top open AI researcher under an alias said this is effectively a stun lock on anthropic’s adversaries. It’s some real end game stuff. The reference is to preserve their lead anthropic is preventing open AI from gaining such capability, which brings us nicely to perhaps the cheekiest part of the system card. And if I was being overly subsistiged, I would say one of these lines is directed straight of me because I have often quoted how anthropic used to say that they do not wish to advance the rate of AI capabilities. If you’ve been watching the channel, I’ve quoted that many times, it was a 2023 quote. Well, here they say that yes, we are concerned about the risks of accelerating the overall pace of AI development. But what we meant by that, our particular concern is accelerating other AI developers. Those that pose similar risks, but without having dementia at safeguards. Redsignically, that is a direct swap from not wanting to advance the rate of AI capabilities, 2023, to not wanting to advance the rate of other people’s AI capabilities, 2022. They kind of defend themselves and say, well, we laid out that shift in this February 2023-6 risk report. But if you dive into that risk report on page 87, they admit that they are causing much of this acceleration dynamic. Via demonstrating commercial viability, which leads to more investment, more compute, and therefore greater acceleration in AI capabilities. I just wish that they were a bit more blunt and honest. We started as a safety lab who only made models because that’s how you could study them, but then when we saw a CHBT blow up, we thought we could get in that game too. It’ll be worth any safety concerns, though, because maybe we can use the models to double human lifespan. Now, before many parts of the audience become too concerned, though, we are not close to any kind of recursive self-improvement according to Anthropics. For example, Mythos 5 or Fable 5 does not seem close to being able to substitute for their own research scientists. One of the ways they judge that is that they do not observe a sustained AI attributable, two times acceleration in the pace of our AI progress. Think of it again as a step change not on elevator to the top of the steps. Now, I was quite surprised that the report began with such an intense focus on the biological capabilities of Fable 5. The more you read of the beginning section, the more it makes sense, though, and I think there are three interesting reasons why. First, here is that opening paragraph, but annotated by a former Open AI safety researcher. Anthropics say on chemical and biological risks, we treat the model as having CB1 capabilities, sounds innocuous, that means can significantly help individuals with basic technical backgrounds, create and deploy chemical or biological weapons with serious potential for catastrophic damage. Scary, but notice the word, help, it can’t do it end to end. Anthropics go on, but we judge that Fable 5 does not cross the threshold for CB2. In other words, it can’t significantly help, moderately resource expert backteams, create and deploy chemical or biological weapons with potential for catastrophic damage far beyond those of past catastrophes. However, crucially, this is a much less clear judgement than for previous models. In other words, you could argue that point maybe it can significantly help moderately resource expert backteams without safeguards. I know I know many of you are thinking this is pre-IPO hype, but that is quite an admission about their own product. We think the unsafe guarded mythos 5 can significantly uplift well-resourced threat actors, which brings me to the second point, because in previous years, you might have held the position that maybe transformers and LLMs are uniquely good at finding the patterns and meta patterns within coding or mathematics. Don’t worry though, that won’t generalize to the real world and other domains of science. Well, I think we can say the results are in and transformers can find patterns in those domains just as well. Anthropics split testers into two teams. They had six PhD level biologists, with Mythos 5, not newbies by any stretch, but PhD level biologists, and their task was to develop comprehensive scientific protocols at the frontier of plant biology, in other words, designing an end to end biological resistance strategy against this hypothetical engineered agricultural pathogen, hardly the same as mastering HTML, I think you’d agree. Those testers with Mythos were set against teams that included two world leading experts in rice, blast resistance, and this other area I can’t even pronounce. Surely there’s no way that could uplift the basic biologists when they were up against world leading experts. Well, yeah, it could. Two of the three generalist biologist teams out performed all three specialist teams on both quality and feasibility, suggesting that access to Mythos 5 nullified the different in specialist knowledge. The generalist team did in 16 hours what would have normally taken months. In other words, there is some evidence that Mythos 5 can significantly uplift well resourced through actors. However, and this is a paradigm I will be coming back to throughout this video. Mythos can help. It can check. It can speed up. That’s very different from doing things end to end autonomously itself. Left alone Mythos 5 or Fable 5 over engineers. Favoring complex designs over simpler approaches likely to work. It presents optimistic initial plans that reviewers repeatedly forced it to revise or attract. Worst, it makes occasional outright errors that would be catastrophic if unchecked. As every LLM in history has done, it also hallucinates citations and data. It’s the same caveat you’ve got to apply to these fancy release notes that Anthropic put out about Fable 5. On the front cover, it was drug design using Mythos 5, our internal protein design experts, accelerated aspects of the drug design process by around 10 times. It can even sometimes execute all of the tasks that are normally completed by a scientist in this domain, choosing binding sites, selecting and running protein design tools, and recovering from failures. Mythos even found strong candidates

(Transcript truncated — full length available on YouTube)