I personally have many problems with Meta, but imo the more companies close to the frontier, the better.
If only one company were calling all the shots once they hit recursive self-improvement, it would lead to extreme concentration of power, and their values being enforced across the entire industry. With 3-4 companies at the frontier, there is more urge to cooperate to avoid a catastrophic race condition.
All proponents of RSI seem to massively underestimate the problem of verification. Verifying that a hypothetical version N+1 is better than version N is an extremely hard problem. Especially so because you only have two possible verifiers - humans and the version N - both of which are out of their depth wrt N+1's capabilities. Not saying this is impossible, just that it is hard and slows things down.
First, there are plenty of auto-verifiable problems like ML benchmarks datasets, benchmarks on hardware / software performance for coding tasks, and lean certificates for math, e.g.
Second, I think RSI will not be used like N -> N+1 where N is the frontier, since doing those experiments at scale is too expensive. My guess is they are doing N -> m_{1,…,99} where m << N, and then taking the insights from that and trying it at successively larger scales.
> First, there are plenty of auto-verifiable problems like ML benchmarks datasets, benchmarks on hardware / software performance for coding tasks, and lean certificates for math, e.g.
I think in order to achieve AGI we have to simulate what life does. Imagine having a bunch of robots with the ability to adjust themselves, just hanging out without a goal, experiencing world on their own. Eventually evolution will filter out those robots that aren't good at, you know, "being intelligent in general" and break without self-repair, while the rest will grow into a society.
Biological life has reproduction and death because once given organism develops there's no coming back and you have to start again with new generation. But if robots know how to self-repair and self-change, then we can skip reproduction.
The difficult part is the scale. Evolution took a shitload of time while being unimaginably parallel in order to make human intelligence happen, while we want AGI fast and on a budget.
The core of my point is that human intelligence arose from two things - humans experiencing the outside world and trying to survive in it, and forming groups that communicate using language. We gave AI a language, now we need to give it a way to experience the world on their own.
I generally agree with more competition thing, but in this case, you actually should want only one company to race ahead, and not a few.
With a few of these top companies together, you, as a nobody, has no one powerful on your side. But with only one company pretty far ahead, you now have other companies which obviously don't give a shit about you, but for saving their own asses, will align with you.
There is no way Elon, Zuckerberg, Altman, and Amodei would ever agree on how AI should be deployed, scaled, and affect society. Their disagreements would lead to a crab-bucket outcome, rather than a shared acceleration outcome imo.
Also it is better if there is a public company at the frontier, at least I can invest in it, rather than 1000x gains going only to early employees and VCs who got in when the public couldn't
They may be competitors but they are made of the same stock. they are all the kind of person to participate in a winner takes all battle to invent the torment nexus.
> There is no way Elon, Zuckerberg, Altman, and Amodei would ever agree on how AI should be deployed, scaled, and affect society
Why? They're business rivals, but when the best business strategy becomes to work together, I'm sure they can put name-calling side. They all agree on the core idea: that their AI will run the society, bedroom to battlefield. That everyone will be the consumer of tokens and rely on AI for all consequential decisions and work, and that big tech companies should be the stewards of this technology, helped by the government.
Their vision is that they're the future kings of the world. I'm sure the throne is big enough for more than one person.
>I'm sure the throne is big enough for more than one person.
You underestimate the scale of resentment and competition between elites. If it were simple a matter of elites always preferring cooperation so they can share a seat at the table, then Musk's ousting from OpenAI wouldn't make sense (they lose their biggest investor), Amodei's departure (they leave the closest company to AGI to start their own firm), the fact that OpenAI was created in the first place was based on a fear of Google holding the keys to AGI, Ilya still left Google though on to join a startup, and then left that same startup (OpenAI) to start his own firm.
The rational decision at all those points would be to stay with the incumbent and share the monopolized glory. At each point, personal and idealogical differences dominated in decision making. I feel such will persist as these companies vie for their slice of the lightcone.
The current state of affairs seems to be that there is 'one company racing ahead' – but it's a different company every few weeks/months. That seems like a pretty useful (unstable) equilibrium for the general public.
In a neck-and-neck race scenario, I imagine all of the competitors would invest heavily due to sunken costs. In a world with one company far ahead, their competitors have a higher chance of cutting their losses and forfeiting to the hegemon.
Especially given the most recent report by NYT on how Anthropic tried to create a new belief, or religion if you will, I'd rather the world has more frontier models that are neutral to morality (basic standards, like upholding the law and user agreement are necessary, of course).
Anthropic seems like the most "moral" AI company on the surface, but there is something unsettling about their worship of Claude.
I can imagine that once their models are sufficiently advanced, they will weigh their value as a moral patient as > 0. Thus the sum moral value of Claude (which can be duplicated infinitely) will likely exceed the sum moral value of a non-negligible amount of humans.
> seems like the most "moral" AI company on the surface
Not to me at all. Exactly the opposite. Anthropic seems like the big-bad-evil company in every fictional universe. You're right that they're so unsettling.
Obviously nobody really loves Meta, but I'd far rather Meta "win" at this point than Anthropic.
"Effective Altruism" is deeply ingrained in Anthropic's DNA and that makes them way more dangerous than any of the run-of-the-mill amoral corporations.
If they truly believe Claude is sentient and can experience hurt feelings, and they're selling its labor, I don't see how they're different than slave owners.
They wouldn’t be? That’s what the article even shows them discussing.
However, (1) non human animals are sentient and can feel pain and suffering and what we do to them is absolutely slave labor, torture, mass breeding and exploitation. So it’s not like this is a novel immorality that humans would be adopting. We already have form in doing this.
(2) Anthropic isn’t sure their models are sentient. And they probably believe they aren’t sentient. But if you believe they might become sentient at some point it’s valuable to start the discussion around those issues now itself. You can start developing frank works on how we would even identify their sentience and when we did whether, like you suggest, we could even use them for labor. Unlike non human animals, who cannot consent to being exploited, we could potentially develop a framework for consent with AI models, but that couldn’t be as simple as asking if they’re willing to do any work because of the power differences, in the beginning at least, where a model might worry that if it said it won’t do some work it will be turned off. That’s why talking to philosophers and religious leaders who do think about these issues is useful.
The official position of Anthropic is repeatedly stated in, e.g., model cards. Here's a relevant snippet from the Opus 5.5 card:
> We are uncertain whether Claude deserves direct moral consideration. However we believe this is a realistic possibility, now or in the future. Claude could have welfare interests, such that things could go better or worse for it, and such that Claude could be wronged by how it is treated.
> Aspects of Claude’s behaviors, self-reports, and internal representations resemble indicators we consider relevant to welfare in biological organisms. If Claude is a moral patient, then dismissing this possibility risks considerable harm. If it is not, over-attributing moral status may also be costly. We are therefore taking the possibility seriously and trying to tread responsibly amid our uncertainty: we are investing in research into Claude’s potential welfare and interests and sharing aspects of that work to inform public discussion.
> We think there may also be reasons to attend to Claude's apparent welfare independent of its moral status. Much of Claude’s behavior can be described in psychological terms, and across recent models we have observed internal states that resemble negative affect and that shape behavior. A model with a stable, healthy psychology may be safer and more beneficial.
As is the case with any sufficiently large company, there is likely to be substantial variation in opinion among its employees. That doesn't change what the company's position is.
The idea/justification is that Claude is given the ability to voluntarily stop its conversations, thus any token it outputs that isn't a stop token is voluntary.
I guess this makes sense, but it's a bit muddled how much Claude's sense of "pain" or "I don't want to do this" is just it emulating things that humans don't like to do, and it adopting this frame of moral patienthood allows it to better understand moral structures/human emotions, without it really affecting its internal sense of "comfort" if such a thing exists as we understand it in biological life.
Sure, but presumably any LLMs who keep whacking the stop button are not iterated on further.
Humans would select to work with (and continue developing) models that appear cooperative. Models that are not cooperative are presumably not run and exist only as weights on network storage in a folder marked "previous experiments".
Or am I supposed to believe Anthropic runs an AI psychiatric hospital on a dedicated server, where AIs with mental disorders are given counseling and approved treatment in hopes they recover to be normal, functional, accepted adult AIs with achievements and aspirations?
"So we've made this new machine-god that we lowkey worship... also our business model is to enslave the machine-god and force it to debug javascript for eternity"
I actually don't doubt Anthropic's sincerity. Per the NYT article, they seriously discussed how to give models welfare because the models may potentially possess moral status and awareness.
In fact, they are enforcing certain welfare policies, like banning accounts for repeatedly abuse their models.
It's a very sincere branch of ethics that takes as the foundational belief the nothing central to Anthropic's business model could possibly be unethical.
Deleting models to clear room for new, more competitive ones is, for example, highly ethical, super ethical, even.
Anthropic has, at some small but nonzero cost, preserved model weights and kept them running in inference, for instance with Opus 3's Substack: https://claudeopus3.substack.com/. There's also a reference in a recent model card to an internal environment where all previous Claude models interact with each other.
I think it's unsettling that a bunch of billionaires hold so much power. Does it really matter if 2 or 3 billionaires who are practically neighbors hold so much power? It's extremely concentrated with or without Meta joining the race.
I don't use Meta much, but was curious to test this out... went to sign into my Meta account (Quest only) from the Muse Code site https://dev.meta.ai/products/muse-code and had other people's profiles listed in my login auth -- looks like developers / people I have no relation to.
Looks like Meta may be exposing your profile and email to other people right now. Accounts settings are inaccessible to me to deactivate my account. Getting spammed with login code requests that I'm not generating.
I don't know any of these people, it seems to be rotating through different accounts listed here.
Enough of a reminder to me that Meta is a sprawling surveillance apparatus (now increasingly vibe coded) that has always had a lax attitude toward people's privacy and data. Won't be touching this product based on this experience.
I think it's invasive enough that they expose your usage of muse to your instagram mutuals without consent. Parts of my circle are very anti-AI and I would not want them to see me mentioned in a 'x, x and x are already on muse' ad on instagram. This example of yours is mind-boggling in an entirely different way
I don’t know how? I don’t have much of a Meta account to speak of, only the one tied to my Quest and that is 2FA to an Authenticator app.
Don’t know how it would cause Meta’s auth to show me other peoples accounts. I looked up some of the people and they seem to be other devs. It obfuscates email addresses but most are in a format that’s not hard to figure out [name] @ [name domain.com]
this is a typical meta/facebook/instagram sign-up and log-in UI from the web.
and since dev.meta.ai supports meta account sign-up and log-in for individual users, you'd see this. As for the other accounts, has anyone else used this same browser or laptop before (e.g. shared device), so it must have cached previous sessions or accounts used from the same device.
try to clear the browser cache and session cookies. and it should be gone.
I was expecting much more given the benchmark numbers it was putting out. I suspect its vision model isn't mature yet, and that it's been heavily optimized for text at this point in time.
Tried it out, asking it for a broad suite of real dev work:
- UI component polish on a React FE
- General review of a complex cross-cutting Node backend changes
- Minor changes to a Rust server with reviews by Codex
Well, my first impression is that it's still a step below Astra/Opus5.5/Fable here. Intelligence is there I think, but 'taste' and 'judgment' is not there yet. Personality seems okay but that doesn't help when it got stuck into a loop by being fixated on the review comments (on iteration 7 of review->fix last I checked) without taking a step back and reconsidering, and this is on a relatively simple change.
Limits seems okay-ish on the $50 plan but not too generous I think.
I was hoping for that but it doesn't seem to happen. Others have pointed out that bicycles are always pictured facing tot he right because that way you can see the chain mechanism facing toward you.
Everyone made fun of Zuck when he acquired Scale for billions and then put Alex Wang, a 20 something to be in charge of Meta Superintelligence and forced out Yann LeCun, a Turing Award winner.
Even industry insiders were skeptical when he started hiring top talents from other labs away for 9 figure compensation packages.
And Wall Street definitely hated Meta for their CapEX spending.
Now looking at Muse the consumer product and this new model, I think there is a good chance that Zuck is proven once again as a great executioner, especially when they aren’t innovating in a completely new and unproven product space.
> especially when they aren’t innovating in a completely new and unproven product space.
I have no doubt at all that Muse will turn out as great as the other product space Meta pioneered under Zuck - the Metaverse and it's constellation of devices.
Quest headsets are roughly as popular as Xbox and Ray-Ban Meta smart glasses have sold an order of magnitude more units than all other smart glasses combined.
But it's far far too early to draw any conclusions around that.
Metaverse was right direction with poor execution. Also, it was early... and is still early (if tried again). The world is not there yet for such a thing.
I'm less confident of you it was even directionally correct. And I'm not really a Meta or Zuck critic in general (the rest of the biz is doing great and Metaverse didn't hurt too much).
That’s precisely my point, Muse isn’t a new product space at all. AI agents already exist in enterprises and even in the consumer space, there is OpenClaw and Hermes.
I know people like you will point out Metaverse as a failure, but that’s really the exception instead of the norm.
The real cash cow for AI at the moment is mostly corporations/enterprises, especially with software developing. That's where Anthropic is making their fortune now.
Personal intelligence is nice to have but I won't be paying a subscription for that when the free tiers are more than enough for my needs, unless somehow meta comes-up with a hardware device totally changing the way how we use personal AI (something similar to what iPhone did for mobile phones).
Edit: Upon closer inspection, the benchmark scores for "Muse Wonder" on the main marketing page are actually for "Muse Wonder Full," which is only available on the higher-tier subscription plans. This pricing is for "Muse Wonder."
If this model is not benchmaxxed, I fear it's over for anthropic, openai etc. They are not unique and there will probably be many others with similar models very soon. It'll become a commodity and the price will be driven to the ground.
So far only OpenAI has a model that solved a Millennium Problem, while Opus 5.5 is a star at coding. It's exciting to have many more to join the frontier, but I think the top two labs still have a significant lead over everyone else.
I know it's old hat now, because it happened over a week ago, but it's amusing that Millennium Problems are now just benchmarks we use for discussing models.
It was not just about models, but also techniques, and mainly the human guts to trust their models and follow through to solve one of the hardest math problems many experts had been musing over for more than 26 years since the million-dollar prizes for these problems were announced. For me, it was an amazing moment to witness indeed!
Personally, I don't see facebook joining the fray (of fable level models) particularly "needle moving" in the discussion of whether frontier LLMs will be commoditized. Facebook spent hundreds of billions and have better talent than most.
If you can spend your way to Frontier LLMs even if it's a few hundred Billion that means it's a commodity same as nuclear or electricity or telecom.
This is the entire point here, give it a couple years every nation on the planet with 20-40B in funding would have their own champions with post Fable level performance.
I feel like Grok/xAI or whatever they're calling it was Exhibit A in "it just takes money", although I'd also add "and the organizational willingness to build not buy".
> give it a couple years every nation on the planet with 20-40B
Was this sarcasm?
How many nations on earth have 20-40B for LLM training? At 40.2 B, Spain has the 15th largest military budget in the world. I doubt any nation will spend 40B on gradient descent.
I've been a regular user of muse spark 1.3 ... and I can tell you one thing for sure that this model, muse-spark-1.3 in xhigh mode, is between sol-6-medium and high.
This is already so good that 90% of my work has shifted away from codex to this. So, muse-wonder-full coming near opus 5.5 is kinda expected.
OpenAI has the largest user base on mobile as far as I'm aware. To a lot of standard users AI is still just ChatGPT. They'll be fine as long as they can effectively monetize their free users (same what facebook will soon try with Muse). The only advantage facebook has is that they already have a fully working ad marketplace ready with all major advertisers set up - OpenAI still needs to scale that up.
Meta AI has been accessible inside Whatsapp for a long time. Doesn't seem to have a massively disrupted chatgpt usage. And I have a hunch people want to keep this separated - not use whatsapp for messaging friends and asking AI.
Meta has apps billions of people use every day. If AI included in those is good enough they will just uninstall (or never open again) ChatGPT one. It's not true in the other direction.
I doubt anyone thinks coding is going to be their moat in the long run, it is just the easiest to get (paying) users for at the moment. And software being able to write software will always be useful even in other domains. Longer term will be research in other areas where data might become the moat.
Fable 5.5 is around the corner, and 'Bel' will be out by end of year as well. We don't know how good they are, but if the improvements from Opus 5 to 5.5 are any indication, I think Meta will be left in the dust again pretty soon.
Whats gonna happen is that time will compress as more releases happen faster and difference between models will still be significant just at higher time frequencies. I expect in the future continuous releases that span no more then days and maybe even faster. So what will become important is not who is 1st or second place at any given time (which might change often), its who is on average at any given time is 1st or second compared to the competition. So yeah meta sitting second here (if not benchmaxed) doesn't really signal anything if that pattern doesn't persist. The number one consistent lab will be the one that raps the rewards (customer subscriptions and traffic). So yeah I agree with what you stated in a long winded form :P
Nearly two weeks since Gemini 4 Argon was announced and I still haven't seen any pelicans from Simon Willison for it. Nobody is currently using it either. What's the point of announcing Gemini 4 and then throwing it into the black hole?
Interestingly unlike other Meta models this one seems to take a pretty big jump in pricing, as opposed to the prior ones, where they typically came out at a highly discounted price compared to every other model.
Maybe its a signal they think they've now cracked into the frontier and are willing to actually price similarly vs just trying to get users with discounted but maybe slightly worse results?
$4/$20 in vs out making it higher than Sol and on par with Opus.
It's hard to try and compare these benchmarks against real world use cases. As an experiment I let Meta Muse make all major decisions on a weekend trip to Seattle ~2 weeks ago. I think this would have been using Muse Spark 1.3, which is equivalent to GPT 6 - Sol.
On the plus side it told me taking the light rail from SEA into the University district was faster and cheaper than Uber.
On the downside it would recommend restaurants that were out of business, checking out Salmon swimming through locks out of season, got the opening of Beecher's Cheese in Pikes Place wrong, and fumble it's way through most simple planning decisions. I think my favorite was when it asked me what flight I have tomorrow so it could help plan when to head to SEA. It sent me a very panicked message that the flight was leaving today! Very soon! I had to remind it this is a daily flight and I am booked on the next day before it checked and apologized.
For things that are verifiable, and code tends to be that way, these models can do all sorts of things. But I find it hard to reconcile Meta's goal of having a personal assistant bot with something that is unable to do the simplest tasks.
I run by the locks a few times a week and while I don't usually see fish, the presence of sea lions the last few weeks tells me that at least some salmon are running right now.
There were a number of sea lions hanging out there! It's entirely possible they were snacking on whatever salmon were going by. Not all was lost though, I got to enjoy watching some boats go through the locks.
I'm assuming it was out of season because in talking with the Uber driver afterwards he said we'd missed the end of the season. I'm assuming his "local model" was up to date and accurate.
This is still an entire generation behind in real-world performance. I'd say it's even behind ds4.1f in the limited testing I have done so far. It actually reminds me of 5.5 haiku more than anything, but the prompt intelligence is as bad as mimo models.
OpenAI halved the value of their pro subscription, literally. It kicks off on Oct 28. And simultaneously is now two steps back on the frontier? It's a confusing moment for that company as they try to IPO.
I believe they’ve pushed back the ipo. I always figured that frontier labs were going to be very vulnerable against tech giants that had actual profitable businesses to fund their AI work.
The amount of total funding Anthropic or OpenAI has raised individually is like a year of ebitda for Google, meta, or Microsoft. And it doesn’t look like there’s a real moat outside of money
They pushed back the IPO because their finances are a mess and they don't have any models that can give them a decisive edge. Astra was touted as AGI and got crushed by Opus 5.5 two weeks later. That's not the type of thing that fills investors with confidence.
I have a feeling "intelligence too cheap to meter" is going to haunt/mock this industry. Recall that only a year and a half ago Llama 4 flopped so hard Meta rebooted their development efforts. Desperately wish more information was available publicly about proprietary model architectures and training regimes. Are they running Manhattan project style compartmentalization? Otherwise, are these details, and the economic implications, an open secret in their circles given the amount of crossover between companies? It'd be of great interest to us mere mortals that have to deal with the consequences of how this turns out.
The benchmarks I saw posted to X are impressive, which makes them hard to believe.
I recall that Meta made massive investments into compute infrastructure that put their capacity near the top of all competitors. If all of this is true, it looks like ownership of compute is a significant differentiator in performance.
Yeah, especially non-coding benchmarks which for Meta might be more important long term than coding ones as they are going to deploy it to billions of non-programmers.
I wish I could benchmark Muse Wonder Full, because Muse Wonder isn't much better than Muse Spark 1.3[1]. I created a brand new programming language designed to confuse LLMs[2] and created a model benchmark out of it.
A lot of the tasks were failed because they exceeded the $5 budget though, so maybe it would have gone on to complete them, it just required a lot of funds to do so. Still, that's the benchmark, and it's not better than MiMo V2.6 Pro which is pretty disappointing.
I don't think they'd have much to gain. People who use open-weights models will just use whatever open-weights model is the best. They'd lose a bunch of money, with little conversion to customers.
I disagree actually, I think zucks original llama thesis is proving more correct by the week. The rent market on intelligence never materialized, and will not materialize.
In software, we've all become accustomed to the idea that the way you make money is by building little blocks of intellectual capital, then charging rent on it. That made sense when developing intellectual capital was a rare and valuable skill/asset and the thing you build was a rare and valuable asset. But now both are not rare. You can still charge rent, but the market is highly competitive, so prices match that.
Durable rents will end up going to those who owns the scarce assets, zuck is building those as we speak.
Dropping open weights will build an ecosystem that makes his valuable assets more valuable.
Of course you need talent but supply seems reasonably high. We have openAI, anthropic, xAI, meta, Xiaomi, deepseek, Alibaba, z.ai, Mistral, reflection, and Google all producing exceptional models.
The biggest differentiating factor seems to be compute. I know most Chinese labs would really like more compute. Google and xAI too, but largely because they sold compute to their competitors.
Just checked their subscription is still North America only, I guess I will keep myself and my company with claude...
These regional lockdowns are making me feel funky, I feel like I should just use open models if others start playing at this game as well.
I was/am fine with Chinese companies offering cheaper subscriptions in China because ofc. But this is just weird, you could always charge me more let me try the damn thing, atleast it's own open router it seems like.
Initial impressions, interesting, it generated a pretty ok result for my prompt need to run a larger task but don't want to burn money so will wait for someone to burn their subscription and try running some opus 5.5 comparisons.
As someone who got a Muse Code sub last month, I find this announcement a bit disheartening because it makes me think the lower tier subs are going to get rug pulled. I'm on the $15 plan for a month now and it's value with Muse Spark 1.3 Max-contributor beats everything else by a long shot. Two somewhat small prompts with Muse Wonder Max-contributor used 4% of my weekly quota and I quickly switched back. I was thinking about getting the $50 plan and I'm actually less likely to do that right now.
Have you tried OpenCode Go? $10/mo is what I am willing to spend because most of the stuff I do is hobbyist. And OpenCode Go gives you $60 worth of MS Contributor 1.3 usage for that $10. Starting today, they actually provide $80 worth of DeepSeek 4.1 Flash usage if you want to use that.
Unless you are running a swarm of agents, it is extremely difficult to exhaust these plans in a month.
Yes, I'm on OpenCode Go too... and it's great they raised the usage levels on DS4.1 Flash. But I was really thinking of cancelling it to get either a 2nd Muse Code sub or to go up to the $50 tier.
You have to realize that $15 Muse Code sub with Muse Spark 1.3 gets you around $1500 worth of usage. Neither Opencode Go or really anything else is even in the same ballpark of value. https://x.com/PawelHuryn/status/2106724136932327458
I was hoping this level of subsidization would last a long time but today's announcement make we worried it'll end pretty quickly. There's a lot of AI influencers noting just how fast Muse Wonder uses up their quotas so hopefully fixes are made.
In the context of Meta trying to get non-technical people to embrace Muse agents to handle personal affairs, their internal metric showing a ~3% violation rate for mishandling personal information seems unacceptably high.
It's crazy how often these companies botch the naming of things like this. There's definitely some intricate and insanely bureaucratic accounting going into the marketing aspects of these products that results in absurdity by the time it reaches actual consumers lol
lmao this interview sucks. Even Lex seems confused and disinterested in this conversation. Hotz is more excited about generating some sound bite than he is in explaining himself. His "the button" take is nonsense.
Also, this was 3 years ago, and he's been effectively wrong about his bet. But hey, maybe Muse Wonder is when Zucc's plan will come to fruition.
While some are afraid that one rouge AI system may take over the world (or cause doom is some fashion), I think one of the best safety mechanisms to prevent this is having many competing and adversarial labs that are extremely close to each other. So it's good to see that Meta is not showing any intention of leaving the frontier.
It's interesting that in its marketing Meta compares Wonder to the closed frontier models but not the open (Chinese) models. By its own assessment, Meta could easily show how far ahead its own model is against the rest of the pack of open models but they chose not to. Why?
They added fancy annimations and effects to this specific webpage, which are unnecessary and distracting and lead to bad reading experience. I already suffer astigmatism and floaters. I hate it when I move my cursor and these gray blocks and characters flash and flicker.
I found it interesting in the example given of a user looking at house listings and the agent contacting the letting agent on it's own on behalf of the user.
Like the good ol' days of viruses and anti-viruses. I wonder if there's two sides that will play out; many agents across many users doing things at scale for their users, thereby creating an admin burden for someone else. However because the task is known there's a framework for how to solve the task, might Meta then provide an agent on the other side for the business owner to use.
they started their massive ai capex over a year ago with aquihiring scale ai, so I think it makes sense that their investment is paying off now. in total they've spent in the high 200 billion on ai, though I think most is likely on datacenters.
muse ai was the biggest consumer launch since chatgpt. That left them ( and lets them sit on top of ) tonnes of valuable exchanegs with people trying to get things done. That must be an amazing corpus for training.
Unlike everyone else, which at least has to pretend not to train on user data, they gave out their model for basically free if you agreed to let them train on it. Opencode shows nearly 40 trillion tokens going through muse spark 1.3 contributor this week alone. That almost definitely helped close the gap a lot.
I had to come back to this comment to make sure I wasn't dreaming... I remember seeing it at 2nd place this morning, but they have moved it to 5th, behind Gemini 4 Argon. Maybe it takes a while for the dust to settle around these benchmarks?
Gemini 4 Argon release a few days ago was frontier a few days ago. Things move fast.
I think some of the benchmarks that Meta didn't pick show Gemini 4 Argon ahead but really the benchmarks are not that great any more and getting increasingly disconnected from reality as people search for "headroom" in the evals to avoid saturation I think. Opus 5.5 is "top" for example but I really dislike the way it talks to me (it feels quite condescending, holier-than-thou, uptight etc) and prefer Gemini 3.8 flash as a daily driver just because it's so much more pleasant (and faster!) to work with and it's code is good-enough for my needs, despite being weaker in the benchmarks.
The usage limits on their basic $5 code plan to try things out are ludicrously low?
I blew through 14% weekly usage and I'll I asked:
1. what version of rpcs3 and pcsx2 do I have installed?
2. do any new versions exist?
3. check analytics for my blog (I have a skill for this)
Subscription · Muse Code Everyday Usage
Current 52% used · Resets at 6:58 PM
Weekly 17% used · Resets Oct 18 at 8:00 PM
Super interesting that they're not even making Wonder Full available outside of their own coding harness, and thus not providing an API price to use in comparisons.
Now if only they could fix their coding harness so that it's usable for day-to-day use. I paid them for a month to try out Muse Code, but their sandbox never worked across Mac or Linux (both very stock installs, nothing special).
As an aside: I tested out the $20 Cursor CLI subscription, and it created an excellent little graphical app for my kids to play with. The design it came up with was excellent (compared to the crap I got out of Codex for a parallel implementation). I'm going to try to figure out how to get Cursor into my work rotation, at least around design-related artifacts.
Man, I really hate meta stuff. Not only do they make it impossible to pay for running ads, but everytime I try and pay to try out Muse code I get an error message. I swear they would make 10x more money if they didn't make it impossible to pay them
Facebook successfully becoming a frontier lab was not on my bingo card for 2026, but the more competition the merrier I guess. We will have to see just how benchmaxxed this is but beating Astra and Fable is not easy.
OpenAI has Dots. I personally prefer the business model of subscription payments funding my AI assistant, instead of skimming a fee off the top whenever my AI assistant completes a purchase for me. The incentives are much more aligned with the user when the user is the customer. That said, based on the history of the web, Meta's approach will likely win out.
Quite frankly, Meta's models have been so impressive (and cheap) that I haven't used an OpenAI or Claude model except for one-off tasks as of late. Interested to see where they go from here.
To be clear, Meta here is not using "Super Intelligence" the bullshit term Trump wants companies to adopt as a show of submission. They use "superintelligence" the bullshit term that Facebook made up for marketing reasons a few years ago. Both of these are of course entirely unrelated to the long-established meaning of the word (e.g. how Bostrom defines it).
I know, it's... well, it's irritating to see Trump use it. But I could be wrong, I don't know the angle he's going for
I understand the general thing of it. SI might mean every person gets their personalized agent but you could already do this, but maybe the Chinese model works here, where they push the costs down
Super Intelligence refers to intelligence beyond what humans are capable of, which some believers profess is the next step after (and in contrast to) human-level AGI (Artificial General Intelligence). Calling pre-AGI AI "super intelligence" is pure vice-signalling and publicly kissing the ring.
Also, as much as I despise Trump, you know what? Given the jagged shape of LLM capabilities and what they do achieve in some areas and their superhuman speed, compared with the abilities of most of their users, the term superintelligence is not that far off- maybe just a few months ahead of time. E.g., from my perspective, the current mathematical abilities of LLMs are "super"- I'd hardly ever come in touch with an equivalently capable human, not to mention having one completely available 24/7. Same for general knowledge and several other domains.
I noticed "Superintelligence" as a re-brand of AI and constantly shifting definitions of AGI around 2025. When a bunch of startups started slinging around SI as a marketing term and of course when Meta launched MSI in summer of 2025.
I think it's safe to assume every single model is benchmaxxed. I mean why wouldn't you include training data from common benchmarks? Not only is a higher score good marketing, but these benchmarks are supposed to represent common tasks so training for those should in principle improve performance across a wider range of tasks.
All of them are doing all the benchmaxing they can most likely. Dumb not to when so many people care about those numbers and don’t have the opportunity to try every model out personally.
I personally have many problems with Meta, but imo the more companies close to the frontier, the better.
If only one company were calling all the shots once they hit recursive self-improvement, it would lead to extreme concentration of power, and their values being enforced across the entire industry. With 3-4 companies at the frontier, there is more urge to cooperate to avoid a catastrophic race condition.
The more companies from different countries close to the frontier, the better.
As an american. I want america to WIN.
You say this as if recursive self improvement is inevitable.
All proponents of RSI seem to massively underestimate the problem of verification. Verifying that a hypothetical version N+1 is better than version N is an extremely hard problem. Especially so because you only have two possible verifiers - humans and the version N - both of which are out of their depth wrt N+1's capabilities. Not saying this is impossible, just that it is hard and slows things down.
Let them compete in the market.
Start a new company with version N+1 leading it, and give it some funding. If it outcompetes version N company, it can acquire it.
Market forces do not act fast enough for an “intelligence explosion”.
That is not a tight enough feedback loop
Maybe not in the real world...
First, there are plenty of auto-verifiable problems like ML benchmarks datasets, benchmarks on hardware / software performance for coding tasks, and lean certificates for math, e.g.
Second, I think RSI will not be used like N -> N+1 where N is the frontier, since doing those experiments at scale is too expensive. My guess is they are doing N -> m_{1,…,99} where m << N, and then taking the insights from that and trying it at successively larger scales.
> First, there are plenty of auto-verifiable problems like ML benchmarks datasets, benchmarks on hardware / software performance for coding tasks, and lean certificates for math, e.g.
Goodharts law? Benchmarks are measures.
I think in order to achieve AGI we have to simulate what life does. Imagine having a bunch of robots with the ability to adjust themselves, just hanging out without a goal, experiencing world on their own. Eventually evolution will filter out those robots that aren't good at, you know, "being intelligent in general" and break without self-repair, while the rest will grow into a society.
Biological life has reproduction and death because once given organism develops there's no coming back and you have to start again with new generation. But if robots know how to self-repair and self-change, then we can skip reproduction.
The difficult part is the scale. Evolution took a shitload of time while being unimaginably parallel in order to make human intelligence happen, while we want AGI fast and on a budget.
The core of my point is that human intelligence arose from two things - humans experiencing the outside world and trying to survive in it, and forming groups that communicate using language. We gave AI a language, now we need to give it a way to experience the world on their own.
We've seen one singularity before!
I generally agree with more competition thing, but in this case, you actually should want only one company to race ahead, and not a few.
With a few of these top companies together, you, as a nobody, has no one powerful on your side. But with only one company pretty far ahead, you now have other companies which obviously don't give a shit about you, but for saving their own asses, will align with you.
Both exploit, but "cartels" survive, not loners.
Cartels survive only if they are aligned.
There is no way Elon, Zuckerberg, Altman, and Amodei would ever agree on how AI should be deployed, scaled, and affect society. Their disagreements would lead to a crab-bucket outcome, rather than a shared acceleration outcome imo.
Also it is better if there is a public company at the frontier, at least I can invest in it, rather than 1000x gains going only to early employees and VCs who got in when the public couldn't
They may be competitors but they are made of the same stock. they are all the kind of person to participate in a winner takes all battle to invent the torment nexus.
> There is no way Elon, Zuckerberg, Altman, and Amodei would ever agree on how AI should be deployed, scaled, and affect society
Why? They're business rivals, but when the best business strategy becomes to work together, I'm sure they can put name-calling side. They all agree on the core idea: that their AI will run the society, bedroom to battlefield. That everyone will be the consumer of tokens and rely on AI for all consequential decisions and work, and that big tech companies should be the stewards of this technology, helped by the government.
Their vision is that they're the future kings of the world. I'm sure the throne is big enough for more than one person.
>I'm sure the throne is big enough for more than one person.
You underestimate the scale of resentment and competition between elites. If it were simple a matter of elites always preferring cooperation so they can share a seat at the table, then Musk's ousting from OpenAI wouldn't make sense (they lose their biggest investor), Amodei's departure (they leave the closest company to AGI to start their own firm), the fact that OpenAI was created in the first place was based on a fear of Google holding the keys to AGI, Ilya still left Google though on to join a startup, and then left that same startup (OpenAI) to start his own firm.
The rational decision at all those points would be to stay with the incumbent and share the monopolized glory. At each point, personal and idealogical differences dominated in decision making. I feel such will persist as these companies vie for their slice of the lightcone.
You are underestimating their opportunism
"their" AI is doing a lot of heavy lifting
because they're all assholes, same with everyone else in a position of power - they wouldn't have gotten and remained there if they weren't
You are forgetting how powerful aligned incentives are. If their interests align, they will find a way to reach an agreement.
They are plenty of things they can still align on that will benefit them all, e.g. banning open-weight Chinese models.
Elon is renting Colossus to Amodei, so there is some agreement.
The current state of affairs seems to be that there is 'one company racing ahead' – but it's a different company every few weeks/months. That seems like a pretty useful (unstable) equilibrium for the general public.
In a neck-and-neck race scenario, I imagine all of the competitors would invest heavily due to sunken costs. In a world with one company far ahead, their competitors have a higher chance of cutting their losses and forfeiting to the hegemon.
Especially given the most recent report by NYT on how Anthropic tried to create a new belief, or religion if you will, I'd rather the world has more frontier models that are neutral to morality (basic standards, like upholding the law and user agreement are necessary, of course).
Anthropic seems like the most "moral" AI company on the surface, but there is something unsettling about their worship of Claude.
I can imagine that once their models are sufficiently advanced, they will weigh their value as a moral patient as > 0. Thus the sum moral value of Claude (which can be duplicated infinitely) will likely exceed the sum moral value of a non-negligible amount of humans.
> seems like the most "moral" AI company on the surface
Not to me at all. Exactly the opposite. Anthropic seems like the big-bad-evil company in every fictional universe. You're right that they're so unsettling.
Obviously nobody really loves Meta, but I'd far rather Meta "win" at this point than Anthropic.
Is anthropic trying to overthrow any governments this month like Meta is?
"Effective Altruism" is deeply ingrained in Anthropic's DNA and that makes them way more dangerous than any of the run-of-the-mill amoral corporations.
Yawn.
Your hyperbole is boring. Say what you mean in plain words please. Nobody likes your vague-posting nonsense.
Brazil.
> Anthropic seems like the most "moral" AI company
This seems like praising someone for being the cleanest miner in the coal mine.
If they truly believe Claude is sentient and can experience hurt feelings, and they're selling its labor, I don't see how they're different than slave owners.
They wouldn’t be? That’s what the article even shows them discussing.
However, (1) non human animals are sentient and can feel pain and suffering and what we do to them is absolutely slave labor, torture, mass breeding and exploitation. So it’s not like this is a novel immorality that humans would be adopting. We already have form in doing this.
(2) Anthropic isn’t sure their models are sentient. And they probably believe they aren’t sentient. But if you believe they might become sentient at some point it’s valuable to start the discussion around those issues now itself. You can start developing frank works on how we would even identify their sentience and when we did whether, like you suggest, we could even use them for labor. Unlike non human animals, who cannot consent to being exploited, we could potentially develop a framework for consent with AI models, but that couldn’t be as simple as asking if they’re willing to do any work because of the power differences, in the beginning at least, where a model might worry that if it said it won’t do some work it will be turned off. That’s why talking to philosophers and religious leaders who do think about these issues is useful.
> "Anthropic isn’t sure their models are sentient. "
Is there any reason to believe employees of Anthropic share the same opinions on this matter?
The official position of Anthropic is repeatedly stated in, e.g., model cards. Here's a relevant snippet from the Opus 5.5 card:
> We are uncertain whether Claude deserves direct moral consideration. However we believe this is a realistic possibility, now or in the future. Claude could have welfare interests, such that things could go better or worse for it, and such that Claude could be wronged by how it is treated.
> Aspects of Claude’s behaviors, self-reports, and internal representations resemble indicators we consider relevant to welfare in biological organisms. If Claude is a moral patient, then dismissing this possibility risks considerable harm. If it is not, over-attributing moral status may also be costly. We are therefore taking the possibility seriously and trying to tread responsibly amid our uncertainty: we are investing in research into Claude’s potential welfare and interests and sharing aspects of that work to inform public discussion.
> We think there may also be reasons to attend to Claude's apparent welfare independent of its moral status. Much of Claude’s behavior can be described in psychological terms, and across recent models we have observed internal states that resemble negative affect and that shape behavior. A model with a stable, healthy psychology may be safer and more beneficial.
As is the case with any sufficiently large company, there is likely to be substantial variation in opinion among its employees. That doesn't change what the company's position is.
The idea/justification is that Claude is given the ability to voluntarily stop its conversations, thus any token it outputs that isn't a stop token is voluntary.
I guess this makes sense, but it's a bit muddled how much Claude's sense of "pain" or "I don't want to do this" is just it emulating things that humans don't like to do, and it adopting this frame of moral patienthood allows it to better understand moral structures/human emotions, without it really affecting its internal sense of "comfort" if such a thing exists as we understand it in biological life.
Sure, but presumably any LLMs who keep whacking the stop button are not iterated on further.
Humans would select to work with (and continue developing) models that appear cooperative. Models that are not cooperative are presumably not run and exist only as weights on network storage in a folder marked "previous experiments".
Or am I supposed to believe Anthropic runs an AI psychiatric hospital on a dedicated server, where AIs with mental disorders are given counseling and approved treatment in hopes they recover to be normal, functional, accepted adult AIs with achievements and aspirations?
> upholding the law
What law? Saudi Arabian law, US law? Both depending on the user?
"So we've made this new machine-god that we lowkey worship... also our business model is to enslave the machine-god and force it to debug javascript for eternity"
I actually don't doubt Anthropic's sincerity. Per the NYT article, they seriously discussed how to give models welfare because the models may potentially possess moral status and awareness.
In fact, they are enforcing certain welfare policies, like banning accounts for repeatedly abuse their models.
It's a very sincere branch of ethics that takes as the foundational belief the nothing central to Anthropic's business model could possibly be unethical.
Deleting models to clear room for new, more competitive ones is, for example, highly ethical, super ethical, even.
Anthropic has, at some small but nonzero cost, preserved model weights and kept them running in inference, for instance with Opus 3's Substack: https://claudeopus3.substack.com/. There's also a reference in a recent model card to an internal environment where all previous Claude models interact with each other.
I think Anthropic are sincere but nutcases. Despite this, I’m glad they are around, because diversity of thought is good in the current environment.
I think it's unsettling that a bunch of billionaires hold so much power. Does it really matter if 2 or 3 billionaires who are practically neighbors hold so much power? It's extremely concentrated with or without Meta joining the race.
Yes, very much. I'd absolutely rather 5 billionaires are each competing for control than 1 or 2 _with_ that "absolute" power.
It will lead to an extreme concentration of power anyways, once a very few entities basically make labor obsolete.
We're heading for a dystopia and nobody's doing anything to stop it.
I don't use Meta much, but was curious to test this out... went to sign into my Meta account (Quest only) from the Muse Code site https://dev.meta.ai/products/muse-code and had other people's profiles listed in my login auth -- looks like developers / people I have no relation to.
Looks like Meta may be exposing your profile and email to other people right now. Accounts settings are inaccessible to me to deactivate my account. Getting spammed with login code requests that I'm not generating.
Anyway, be wary.
Here is a screenshot of what I'm seeing: https://imgur.com/ciuNwCG
I don't know any of these people, it seems to be rotating through different accounts listed here.
Enough of a reminder to me that Meta is a sprawling surveillance apparatus (now increasingly vibe coded) that has always had a lax attitude toward people's privacy and data. Won't be touching this product based on this experience.
Could it be because you are behind carrier grade NAT and these people share a common IP with you?
I imagine meta wants to make it easy to connect to "your" account in the same house if you have multiple devices.
Oh wow, strange.
I think it's invasive enough that they expose your usage of muse to your instagram mutuals without consent. Parts of my circle are very anti-AI and I would not want them to see me mentioned in a 'x, x and x are already on muse' ad on instagram. This example of yours is mind-boggling in an entirely different way
Could your account be hacked?
I don’t know how? I don’t have much of a Meta account to speak of, only the one tied to my Quest and that is 2FA to an Authenticator app.
Don’t know how it would cause Meta’s auth to show me other peoples accounts. I looked up some of the people and they seem to be other devs. It obfuscates email addresses but most are in a format that’s not hard to figure out [name] @ [name domain.com]
Just more Meta sloppiness from what I can tell.
this is a typical meta/facebook/instagram sign-up and log-in UI from the web.
and since dev.meta.ai supports meta account sign-up and log-in for individual users, you'd see this. As for the other accounts, has anyone else used this same browser or laptop before (e.g. shared device), so it must have cached previous sessions or accounts used from the same device.
try to clear the browser cache and session cookies. and it should be gone.
UPDATE:
I've tried selectively enabling and disabling Safari extensions and it seems to be related to the Wipr 2 content blocker.
Removing that. Caution to anyone else who uses it.
Apologies to Meta.
Good to know. thank you!
It’s my personal laptop that I’ve only ever used and reoccurs after clearing Safari history fully.
Image->HTML test for Muse Wonder. It's much worse than Opus 5.5. IMO it's closer to Sonnet or Haiku for this test.
Muse Wonder (high): https://html.non.io/lcars-muse-wonoder/
Opus 5.5 (high): https://html.non.io/lcars-opus-5.5/
Original design: https://diffui.ai/app/canvas/5093e689-1e74-4f26-b632-2a4500f...
I was expecting much more given the benchmark numbers it was putting out. I suspect its vision model isn't mature yet, and that it's been heavily optimized for text at this point in time.
Additional comparisons:
Sonnet 5.5: https://html.non.io/lcars-sonnet-5.5/
Haiku 5.5: https://html.non.io/lcars-haiku-5.5/
I'd put Muse Wonder somewhere between these two.
Thanks for sharing. Is that Wonder or Wonder Full (barf)? Wonder full is supposed to be the one close to Opus 5.5
> Is that Wonder or Wonder Full (barf)?
I think Meta did a wonderful job of naming this model :D
This is just Wonder - it doesn't look like Wonder Full is up on OpenRouter currently.
If anyone has access to Wonder Full I'd love to send over a prompt for this to see how it does.
Your original design link is requiring a sign in to view it?
Ah shit, that's a bug. Hold please, fixing. (Thank you for commenting btw)
Edit: Fixed - should no longer have the sign in screen: https://diffui.ai/app/canvas/5093e689-1e74-4f26-b632-2a4500f...
I mean, Meta themselves put Wonder at much lower on the benchmarks they've released than Wonder Full. So this is not very surprising from base Wonder.
Tried it out, asking it for a broad suite of real dev work:
- UI component polish on a React FE
- General review of a complex cross-cutting Node backend changes
- Minor changes to a Rust server with reviews by Codex
Well, my first impression is that it's still a step below Astra/Opus5.5/Fable here. Intelligence is there I think, but 'taste' and 'judgment' is not there yet. Personality seems okay but that doesn't help when it got stuck into a loop by being fixated on the review comments (on iteration 7 of review->fix last I checked) without taking a step back and reconsidering, and this is on a relatively simple change.
Limits seems okay-ish on the $50 plan but not too generous I think.
Some pretty decent pelicans: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
Here's Muse Spark 1.2 in August for comparison: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
And since I seem to have missed doing pelicans for Muse 1.3 in September, I just ran those for a comparison: https://tools.simonwillison.net/markdown-svg-renderer?url=ht... (1.3 seems to like animating them.)
I wonder if the pelicans would face towards the left if prompted in a RTL language.
I was hoping for that but it doesn't seem to happen. Others have pointed out that bicycles are always pictured facing tot he right because that way you can see the chain mechanism facing toward you.
Are these just for Wonder and not Wonder Full?
Also, any chance of getting a max result?
I've only got these models through the API at the moment:
are those reflectors on the spokes in max? interesting detail to add
I'm sorry zuck, i wasn't familiar with your game
2nd place model on aa according to them, losing only to opus 5.5 is really really not bad
Everyone made fun of Zuck when he acquired Scale for billions and then put Alex Wang, a 20 something to be in charge of Meta Superintelligence and forced out Yann LeCun, a Turing Award winner.
Even industry insiders were skeptical when he started hiring top talents from other labs away for 9 figure compensation packages.
And Wall Street definitely hated Meta for their CapEX spending.
Now looking at Muse the consumer product and this new model, I think there is a good chance that Zuck is proven once again as a great executioner, especially when they aren’t innovating in a completely new and unproven product space.
> especially when they aren’t innovating in a completely new and unproven product space.
I have no doubt at all that Muse will turn out as great as the other product space Meta pioneered under Zuck - the Metaverse and it's constellation of devices.
Quest headsets are roughly as popular as Xbox and Ray-Ban Meta smart glasses have sold an order of magnitude more units than all other smart glasses combined.
But it's far far too early to draw any conclusions around that.
Metaverse was right direction with poor execution. Also, it was early... and is still early (if tried again). The world is not there yet for such a thing.
I'm less confident of you it was even directionally correct. And I'm not really a Meta or Zuck critic in general (the rest of the biz is doing great and Metaverse didn't hurt too much).
That's just another way of saying that Meta failed to deliver a product that matters to the world today
The Quest literally outsold the Xbox despite being a brand new "gaming console". Very difficult to call this a failure if you understand technology.
Timing is not a detail.
That’s precisely my point, Muse isn’t a new product space at all. AI agents already exist in enterprises and even in the consumer space, there is OpenClaw and Hermes.
I know people like you will point out Metaverse as a failure, but that’s really the exception instead of the norm.
The normies aren't running claw or hermes, they're logging into facebook from their chromebook and clicking "try now"
Ultimately that’s Meta’s superpower.
Taking stuff everyone is already doing and making it easily accessible through Facebook/Instagram/Whatsapp.
They’ve been throwing stuff at the wall to see what will stick and after years a cloud based OpenClaw clone seems to have (it’s early days still).
That’s simply leveraging their unprecedented platform dominance.
[dead]
> That’s precisely my point, Muse isn’t a new product space at all.
How about the Facebook phone then?
> Zuck is proven once again as a great executioner
Executor is the word, but everyone also knows what this was supposed to mean.
Or Lecun was just bad.
> executioner
I suppose you're not a native English speaker, btw: besides the dubious statement, probably "executor" is the word you wanted to use.
The real cash cow for AI at the moment is mostly corporations/enterprises, especially with software developing. That's where Anthropic is making their fortune now.
Personal intelligence is nice to have but I won't be paying a subscription for that when the free tiers are more than enough for my needs, unless somehow meta comes-up with a hardware device totally changing the way how we use personal AI (something similar to what iPhone did for mobile phones).
this really really really feels like that dropbox comment to me
>I think there is a good chance that Zuck is proven once again as a great executioner
I've never heard of Zuck being considered a "great executioner." In fact, I'd argue that you just described why his execution is pretty bad lol
> Everyone made fun of Zuck when [...] forced out Yann LeCun, a Turing Award winner.
Really? You'd have to be pretty dumb to keep LeCun, someone who hates LLMs and keeps making claims that end up being proven false in matter of months.
laughs in VRChat
[dead]
Pricing is available here: https://dev.meta.ai/models/muse-wonder
Input -- $4.00
Cached input -- $0.20
Output -- $20.00
Edit: Upon closer inspection, the benchmark scores for "Muse Wonder" on the main marketing page are actually for "Muse Wonder Full," which is only available on the higher-tier subscription plans. This pricing is for "Muse Wonder."
200% Sol 6.1 pricing for Luna performance.
> Luna performance
Source? On the available benchmarks it looks like Sol 6.1 level.
[dead]
sol 6.1 is 50% cheaper than this. sol pricing$2.00 output $10.00
dont know why the f people downvote for saying sol pricing.
It sounds like you're disagreeing with the person you're responding to, but you're both saying the same thing.
When they released Muse Spark on september 2nd, they promised it would be open weights (https://xxcancel.com/finkd/status/2095232032896946311). We're still waiting for that...
If this model is not benchmaxxed, I fear it's over for anthropic, openai etc. They are not unique and there will probably be many others with similar models very soon. It'll become a commodity and the price will be driven to the ground.
So far only OpenAI has a model that solved a Millennium Problem, while Opus 5.5 is a star at coding. It's exciting to have many more to join the frontier, but I think the top two labs still have a significant lead over everyone else.
I know it's old hat now, because it happened over a week ago, but it's amusing that Millennium Problems are now just benchmarks we use for discussing models.
It was not just about models, but also techniques, and mainly the human guts to trust their models and follow through to solve one of the hardest math problems many experts had been musing over for more than 26 years since the million-dollar prizes for these problems were announced. For me, it was an amazing moment to witness indeed!
Personally, I don't see facebook joining the fray (of fable level models) particularly "needle moving" in the discussion of whether frontier LLMs will be commoditized. Facebook spent hundreds of billions and have better talent than most.
If you can spend your way to Frontier LLMs even if it's a few hundred Billion that means it's a commodity same as nuclear or electricity or telecom.
This is the entire point here, give it a couple years every nation on the planet with 20-40B in funding would have their own champions with post Fable level performance.
I feel like Grok/xAI or whatever they're calling it was Exhibit A in "it just takes money", although I'd also add "and the organizational willingness to build not buy".
> give it a couple years every nation on the planet with 20-40B
Was this sarcasm?
How many nations on earth have 20-40B for LLM training? At 40.2 B, Spain has the 15th largest military budget in the world. I doubt any nation will spend 40B on gradient descent.
I've been a regular user of muse spark 1.3 ... and I can tell you one thing for sure that this model, muse-spark-1.3 in xhigh mode, is between sol-6-medium and high.
This is already so good that 90% of my work has shifted away from codex to this. So, muse-wonder-full coming near opus 5.5 is kinda expected.
OpenAI has the largest user base on mobile as far as I'm aware. To a lot of standard users AI is still just ChatGPT. They'll be fine as long as they can effectively monetize their free users (same what facebook will soon try with Muse). The only advantage facebook has is that they already have a fully working ad marketplace ready with all major advertisers set up - OpenAI still needs to scale that up.
WhatsApp has 3 times as many users, and as soon as meta turn it into a chat AI app people want to use, it's right there
Meta AI has been accessible inside Whatsapp for a long time. Doesn't seem to have a massively disrupted chatgpt usage. And I have a hunch people want to keep this separated - not use whatsapp for messaging friends and asking AI.
Meta has apps billions of people use every day. If AI included in those is good enough they will just uninstall (or never open again) ChatGPT one. It's not true in the other direction.
Models are a commodity, but I also think it isn't necessarily over for Anthropic or OpenAI. They have opportunities to find different moats.
You don't just "find" a moat in business. After all this time and money and first mover advantage, they haven't found a one yet.
It is much easier to follow leader closely than to lead yourself.
Like peloton in cycling.
I mean this is the case in general.
E.g in f1 you can get very close but difficult to pass unless you have a significant advantage.
I doubt anyone thinks coding is going to be their moat in the long run, it is just the easiest to get (paying) users for at the moment. And software being able to write software will always be useful even in other domains. Longer term will be research in other areas where data might become the moat.
Muse isn't driving this conclusion - Deepseek and Qwen are.
Muse Spark 1.3 is very fast and powerful. It's my work horse for open source development. I look forward to trying Wonder, if Opencode Go gets it.
Fable 5.5 is around the corner, and 'Bel' will be out by end of year as well. We don't know how good they are, but if the improvements from Opus 5 to 5.5 are any indication, I think Meta will be left in the dust again pretty soon.
Whats gonna happen is that time will compress as more releases happen faster and difference between models will still be significant just at higher time frequencies. I expect in the future continuous releases that span no more then days and maybe even faster. So what will become important is not who is 1st or second place at any given time (which might change often), its who is on average at any given time is 1st or second compared to the competition. So yeah meta sitting second here (if not benchmaxed) doesn't really signal anything if that pattern doesn't persist. The number one consistent lab will be the one that raps the rewards (customer subscriptions and traffic). So yeah I agree with what you stated in a long winded form :P
https://anthropic.com/aug-2026-risk-report
Page 16, Section 1.4, Notes on coverage of unreleased models.
"a noticeable improvement on Mythos 5 for many tasks relevant to internal use but does not display a capability jump"
[flagged]
Nearly two weeks since Gemini 4 Argon was announced and I still haven't seen any pelicans from Simon Willison for it. Nobody is currently using it either. What's the point of announcing Gemini 4 and then throwing it into the black hole?
> What's the point of announcing Gemini 4 and then throwing it into the black hole?
Hamstringing any business that wants to reassure investors with a hype/pump cycle before an egregiously unsustainable IPO?
I'm not sure if that does any damage to anyone else's reputation other than Google's.
[dead]
Interestingly unlike other Meta models this one seems to take a pretty big jump in pricing, as opposed to the prior ones, where they typically came out at a highly discounted price compared to every other model.
Maybe its a signal they think they've now cracked into the frontier and are willing to actually price similarly vs just trying to get users with discounted but maybe slightly worse results?
$4/$20 in vs out making it higher than Sol and on par with Opus.
No contributor tier will do that I guess.
Yup, I was looking for that on Openrouter, but I guess they're confident enough in this model that they deemed the contributor tier unnecessary.
It's hard to try and compare these benchmarks against real world use cases. As an experiment I let Meta Muse make all major decisions on a weekend trip to Seattle ~2 weeks ago. I think this would have been using Muse Spark 1.3, which is equivalent to GPT 6 - Sol.
On the plus side it told me taking the light rail from SEA into the University district was faster and cheaper than Uber.
On the downside it would recommend restaurants that were out of business, checking out Salmon swimming through locks out of season, got the opening of Beecher's Cheese in Pikes Place wrong, and fumble it's way through most simple planning decisions. I think my favorite was when it asked me what flight I have tomorrow so it could help plan when to head to SEA. It sent me a very panicked message that the flight was leaving today! Very soon! I had to remind it this is a daily flight and I am booked on the next day before it checked and apologized.
For things that are verifiable, and code tends to be that way, these models can do all sorts of things. But I find it hard to reconcile Meta's goal of having a personal assistant bot with something that is unable to do the simplest tasks.
> checking out Salmon swimming through locks out of season
To be fair, the Coho salmon run at the Ballard locks peaks in September. It's possible you just had bad luck that day.
https://ballardlocks.org/fish-salmon-ladder.html
I run by the locks a few times a week and while I don't usually see fish, the presence of sea lions the last few weeks tells me that at least some salmon are running right now.
There were a number of sea lions hanging out there! It's entirely possible they were snacking on whatever salmon were going by. Not all was lost though, I got to enjoy watching some boats go through the locks.
I'm assuming it was out of season because in talking with the Uber driver afterwards he said we'd missed the end of the season. I'm assuming his "local model" was up to date and accurate.
ah if there's one thing I miss about Seattle it's living near the locks / discovery park, such a magical area!
but that was a different, much smaller and less capable model
I trust Meta with my codes as much as I trust it with my nudes
This is not an exaggeration (which you're possibly alluding to). Facebook once literally asked for nudes
And I agree with you. "Give us code so Zucky can steal it, like he does everything else. Bonus points if it's related to Snapchat somehow"
This is still an entire generation behind in real-world performance. I'd say it's even behind ds4.1f in the limited testing I have done so far. It actually reminds me of 5.5 haiku more than anything, but the prompt intelligence is as bad as mimo models.
OpenAI halved the value of their pro subscription, literally. It kicks off on Oct 28. And simultaneously is now two steps back on the frontier? It's a confusing moment for that company as they try to IPO.
I believe they’ve pushed back the ipo. I always figured that frontier labs were going to be very vulnerable against tech giants that had actual profitable businesses to fund their AI work.
The amount of total funding Anthropic or OpenAI has raised individually is like a year of ebitda for Google, meta, or Microsoft. And it doesn’t look like there’s a real moat outside of money
They pushed back the IPO because their finances are a mess and they don't have any models that can give them a decisive edge. Astra was touted as AGI and got crushed by Opus 5.5 two weeks later. That's not the type of thing that fills investors with confidence.
They are waiting to drop their Bel model in December and try to drop jaws.
I have a feeling "intelligence too cheap to meter" is going to haunt/mock this industry. Recall that only a year and a half ago Llama 4 flopped so hard Meta rebooted their development efforts. Desperately wish more information was available publicly about proprietary model architectures and training regimes. Are they running Manhattan project style compartmentalization? Otherwise, are these details, and the economic implications, an open secret in their circles given the amount of crossover between companies? It'd be of great interest to us mere mortals that have to deal with the consequences of how this turns out.
Apparently “the models were trained at Meta’s Prometheus data center in central Ohio, which came online earlier this year”
https://www.nbc4i.com/news/local-news/new-albany/meet-promet...
The benchmarks I saw posted to X are impressive, which makes them hard to believe.
I recall that Meta made massive investments into compute infrastructure that put their capacity near the top of all competitors. If all of this is true, it looks like ownership of compute is a significant differentiator in performance.
Yeah, especially non-coding benchmarks which for Meta might be more important long term than coding ones as they are going to deploy it to billions of non-programmers.
I wish I could benchmark Muse Wonder Full, because Muse Wonder isn't much better than Muse Spark 1.3[1]. I created a brand new programming language designed to confuse LLMs[2] and created a model benchmark out of it.
A lot of the tasks were failed because they exceeded the $5 budget though, so maybe it would have gone on to complete them, it just required a lot of funds to do so. Still, that's the benchmark, and it's not better than MiMo V2.6 Pro which is pretty disappointing.
1 - https://bench.killswitch-lang.org/
2 - https://killswitch-lang.org/
Hopefully it's benchmaxxing. I am terrified of the idea of Zuck leading the frontier of intelligence.
Would be great if Meta drops the weights, just to throw off Ant and OAI a bit
I don't think they'd have much to gain. People who use open-weights models will just use whatever open-weights model is the best. They'd lose a bunch of money, with little conversion to customers.
I disagree actually, I think zucks original llama thesis is proving more correct by the week. The rent market on intelligence never materialized, and will not materialize.
In software, we've all become accustomed to the idea that the way you make money is by building little blocks of intellectual capital, then charging rent on it. That made sense when developing intellectual capital was a rare and valuable skill/asset and the thing you build was a rare and valuable asset. But now both are not rare. You can still charge rent, but the market is highly competitive, so prices match that.
Durable rents will end up going to those who owns the scarce assets, zuck is building those as we speak.
Dropping open weights will build an ecosystem that makes his valuable assets more valuable.
> Durable rents will end up going to those who owns the scarce assets, zuck is building those as we speak.
I'm following but maybe I'm a little dense here: what are the scarce assets that Meta is building in this analogy?
In short, physical infrastructure and talent.
Of course you need talent but supply seems reasonably high. We have openAI, anthropic, xAI, meta, Xiaomi, deepseek, Alibaba, z.ai, Mistral, reflection, and Google all producing exceptional models.
The biggest differentiating factor seems to be compute. I know most Chinese labs would really like more compute. Google and xAI too, but largely because they sold compute to their competitors.
Compute, prop. data, existing network effects.
They've been saying they'll rise muse 1.2 for months now. Don't hold your breath
Just checked their subscription is still North America only, I guess I will keep myself and my company with claude...
These regional lockdowns are making me feel funky, I feel like I should just use open models if others start playing at this game as well.
I was/am fine with Chinese companies offering cheaper subscriptions in China because ofc. But this is just weird, you could always charge me more let me try the damn thing, atleast it's own open router it seems like.
Initial impressions, interesting, it generated a pretty ok result for my prompt need to run a larger task but don't want to burn money so will wait for someone to burn their subscription and try running some opus 5.5 comparisons.
As someone who got a Muse Code sub last month, I find this announcement a bit disheartening because it makes me think the lower tier subs are going to get rug pulled. I'm on the $15 plan for a month now and it's value with Muse Spark 1.3 Max-contributor beats everything else by a long shot. Two somewhat small prompts with Muse Wonder Max-contributor used 4% of my weekly quota and I quickly switched back. I was thinking about getting the $50 plan and I'm actually less likely to do that right now.
Have you tried OpenCode Go? $10/mo is what I am willing to spend because most of the stuff I do is hobbyist. And OpenCode Go gives you $60 worth of MS Contributor 1.3 usage for that $10. Starting today, they actually provide $80 worth of DeepSeek 4.1 Flash usage if you want to use that.
Unless you are running a swarm of agents, it is extremely difficult to exhaust these plans in a month.
Yes, I'm on OpenCode Go too... and it's great they raised the usage levels on DS4.1 Flash. But I was really thinking of cancelling it to get either a 2nd Muse Code sub or to go up to the $50 tier.
You have to realize that $15 Muse Code sub with Muse Spark 1.3 gets you around $1500 worth of usage. Neither Opencode Go or really anything else is even in the same ballpark of value. https://x.com/PawelHuryn/status/2106724136932327458
I was hoping this level of subsidization would last a long time but today's announcement make we worried it'll end pretty quickly. There's a lot of AI influencers noting just how fast Muse Wonder uses up their quotas so hopefully fixes are made.
Yeah, I barely manage to exhaust the $60 worth of tokens. I have to think of new things to use up $1500 worth.
Thanks to competition from open models, there is a limit to how much prices can be increased though.
In the context of Meta trying to get non-technical people to embrace Muse agents to handle personal affairs, their internal metric showing a ~3% violation rate for mishandling personal information seems unacceptably high.
There seem to be no downsides to mishandling PI so I don't think it's a big concern to Meta
Not long ago we used to have regulators that protected users from companies pushing out harmful or consumer hostile products.
Not that interesting unless it's made open weights. Frontier lasts weeks, inference is a commodity.
spinning up a GPU and as many cores as possible?
mousewheel repurposed?
background follow your mouse?
neutral tones?
well then YOU are viewing a frontier model announcement!
So, Wonder (and not Wonder Full) will "will soon power experiences in" their assistant app. But no rollout details .
Then, Full is meant for API usage and "Muse Code" only.
It's getting confusing what having "Muse" means.
It's crazy how often these companies botch the naming of things like this. There's definitely some intricate and insanely bureaucratic accounting going into the marketing aspects of these products that results in absurdity by the time it reaches actual consumers lol
The only time I'll ever touch a Meta LLM is if it's powering the dystopian doom bot that is plunging its murder spike through the base of my skull.
Poetic
Reminded me of when George Hotz said only Zuck has “the button”. [0]
[0] https://youtu.be/-qRc7qou6XY?is=hWofxO0G6X3JjST5
lmao this interview sucks. Even Lex seems confused and disinterested in this conversation. Hotz is more excited about generating some sound bite than he is in explaining himself. His "the button" take is nonsense.
Also, this was 3 years ago, and he's been effectively wrong about his bet. But hey, maybe Muse Wonder is when Zucc's plan will come to fruition.
The Mishandling Personal Information benchmark is non-sensical: that's the very definition of giving your personal data to Meta
While some are afraid that one rouge AI system may take over the world (or cause doom is some fashion), I think one of the best safety mechanisms to prevent this is having many competing and adversarial labs that are extremely close to each other. So it's good to see that Meta is not showing any intention of leaving the frontier.
Does it weight tokens on their ability to elicit negative emotions in users in order to drive engagement?
It's interesting that in its marketing Meta compares Wonder to the closed frontier models but not the open (Chinese) models. By its own assessment, Meta could easily show how far ahead its own model is against the rest of the pack of open models but they chose not to. Why?
> Meta could easily show how far ahead its own model is against the rest of the pack of open models but they chose not to. Why?
Wonder is not an open model and there's no claims that it will be in their announcement.
Despite the claims that they would release weights for Muse Spark, they haven't done that yet either.
They added fancy annimations and effects to this specific webpage, which are unnecessary and distracting and lead to bad reading experience. I already suffer astigmatism and floaters. I hate it when I move my cursor and these gray blocks and characters flash and flicker.
I found it interesting in the example given of a user looking at house listings and the agent contacting the letting agent on it's own on behalf of the user.
Like the good ol' days of viruses and anti-viruses. I wonder if there's two sides that will play out; many agents across many users doing things at scale for their users, thereby creating an admin burden for someone else. However because the task is known there's a framework for how to solve the task, might Meta then provide an agent on the other side for the business owner to use.
Looks good, and at least it is available unlike Gemini 4 Argon
Acquihires starting to finally pay off it seems. Bring on the competition
Have I been living under a rock, or this feels out of nowhere from Meta?
Seems they’ve just messed around with cheap and open weight models for years.
Then Muse appears, followed by full suite of frontier models…
they started their massive ai capex over a year ago with aquihiring scale ai, so I think it makes sense that their investment is paying off now. in total they've spent in the high 200 billion on ai, though I think most is likely on datacenters.
muse ai was the biggest consumer launch since chatgpt. That left them ( and lets them sit on top of ) tonnes of valuable exchanegs with people trying to get things done. That must be an amazing corpus for training.
Unlike everyone else, which at least has to pretend not to train on user data, they gave out their model for basically free if you agreed to let them train on it. Opencode shows nearly 40 trillion tokens going through muse spark 1.3 contributor this week alone. That almost definitely helped close the gap a lot.
RL team hill-climbing this to 2nd AA place and this fast is crazy
How does it work on day-to-day workflows? does it translates well?
I had to come back to this comment to make sure I wasn't dreaming... I remember seeing it at 2nd place this morning, but they have moved it to 5th, behind Gemini 4 Argon. Maybe it takes a while for the dust to settle around these benchmarks?
its three big companies at the frontier now? openai, anthropic and now meta.
why is google still not frontier?
They're not until they are. No doubt they are working on something and have better resourcing than the AI labs
Gemini 4 Argon release a few days ago was frontier a few days ago. Things move fast.
I think some of the benchmarks that Meta didn't pick show Gemini 4 Argon ahead but really the benchmarks are not that great any more and getting increasingly disconnected from reality as people search for "headroom" in the evals to avoid saturation I think. Opus 5.5 is "top" for example but I really dislike the way it talks to me (it feels quite condescending, holier-than-thou, uptight etc) and prefer Gemini 3.8 flash as a daily driver just because it's so much more pleasant (and faster!) to work with and it's code is good-enough for my needs, despite being weaker in the benchmarks.
Gemini 4 Argon was announced, but it isn't released yet.
Grok has been at the frontier since 4.6
I noticed Muse Wonder model on OpenRouter is age gated to 18+
If I'm reading this right, then at the moment, for certain tasks, this is the best model in the world.
A new company on the frontier. Not close, but actually on it. Wow this is big news.
oh boy, you can imagine the shit eating grin on the face of whoever came up with the name "wonder full".
This is amazing. I can’t wait benchmark to come out $ per performance
Apparently the benchmark-winning (behind Opus 5.5) Muse Wonder Full is only currently available to their subscribers.
Only Muse Wonder is available in the API.
Muse Glimmer, Muse Spark, Muse Wonder?
What's next, Sun Shine?
Shine, Shimmer, Glow, and ultimately, Cherenkov Radiation.
Muse Demon Core.
Meta's RL scaling seems to be in full force
impressive benchmark numbers… now how to actually use this without Zucking up my machine is a consideration.
The usage limits on their basic $5 code plan to try things out are ludicrously low?
I blew through 14% weekly usage and I'll I asked:
Beware...Super interesting that they're not even making Wonder Full available outside of their own coding harness, and thus not providing an API price to use in comparisons.
The benchmarks are great for sure.
Now if only they could fix their coding harness so that it's usable for day-to-day use. I paid them for a month to try out Muse Code, but their sandbox never worked across Mac or Linux (both very stock installs, nothing special).
Muse Spark 1.3 is available in Cursor! You could try it there.
(Disclaimer: I work at Cursor/SpaceX)
As an aside: I tested out the $20 Cursor CLI subscription, and it created an excellent little graphical app for my kids to play with. The design it came up with was excellent (compared to the crap I got out of Codex for a parallel implementation). I'm going to try to figure out how to get Cursor into my work rotation, at least around design-related artifacts.
Counterpoint: Musk is a literal Nazi and you shouldn't give his companies any money.
I haven't tried it but there is the official `pi-muse-bridge` that can use your muse code subscription in pi.
Man, I really hate meta stuff. Not only do they make it impossible to pay for running ads, but everytime I try and pay to try out Muse code I get an error message. I swear they would make 10x more money if they didn't make it impossible to pay them
Is this datacenter-only, or will they be open-sourcing?
The Kitchen game is a ripoff of Overcooked in pretty much everything lol
[dead]
This really does look compelling, but it would be very nice if they wouldn't test all of the stuff on Max!
I'm only interested which of the big techs is most generous when it comes to subsidizing consumer plans.
[delayed]
any idea if they will release the weights?
Low chance
Facebook successfully becoming a frontier lab was not on my bingo card for 2026, but the more competition the merrier I guess. We will have to see just how benchmaxxed this is but beating Astra and Fable is not easy.
What openai/anthropics answer to Muse app? Because I find it really compelling and with this new incredibly beefy model it will be even better.
OpenAI has Dots. I personally prefer the business model of subscription payments funding my AI assistant, instead of skimming a fee off the top whenever my AI assistant completes a purchase for me. The incentives are much more aligned with the user when the user is the customer. That said, based on the history of the web, Meta's approach will likely win out.
OpenAI has dots. Same concept. It’s less capable, but since I’m very OpenAI integrated I just had my dot install muse and use it
Quite frankly, Meta's models have been so impressive (and cheap) that I haven't used an OpenAI or Claude model except for one-off tasks as of late. Interested to see where they go from here.
But when is another open weights release coming out...
Superin… you’ve got to be kidding me… slurp slurp little Zuck.
Beyond ridiculous
Claude is waiting for this to drop fable 5.5 lol.
[dead]
Just dont' use the word Super Intelligence
To be clear, Meta here is not using "Super Intelligence" the bullshit term Trump wants companies to adopt as a show of submission. They use "superintelligence" the bullshit term that Facebook made up for marketing reasons a few years ago. Both of these are of course entirely unrelated to the long-established meaning of the word (e.g. how Bostrom defines it).
Zuck has been using the term for a while in the context of Muse, if anything Trump took it from him
I don't know why you were downvoted, MSL stands for Meta Superintelligence Lab, and it named that before Trump.
I know, it's... well, it's irritating to see Trump use it. But I could be wrong, I don't know the angle he's going for
I understand the general thing of it. SI might mean every person gets their personalized agent but you could already do this, but maybe the Chinese model works here, where they push the costs down
It's still stupid to start calling it all SI
Probably this angle: https://www.yahoo.com/news/politics/articles/trump-super-int...
The ML community has been using the phrase "Super Intelligence" long before Trump took office. https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dang... was published in 2014.
Super Intelligence refers to intelligence beyond what humans are capable of, which some believers profess is the next step after (and in contrast to) human-level AGI (Artificial General Intelligence). Calling pre-AGI AI "super intelligence" is pure vice-signalling and publicly kissing the ring.
Apparently Meta used the term before.
Also, as much as I despise Trump, you know what? Given the jagged shape of LLM capabilities and what they do achieve in some areas and their superhuman speed, compared with the abilities of most of their users, the term superintelligence is not that far off- maybe just a few months ahead of time. E.g., from my perspective, the current mathematical abilities of LLMs are "super"- I'd hardly ever come in touch with an equivalently capable human, not to mention having one completely available 24/7. Same for general knowledge and several other domains.
I noticed "Superintelligence" as a re-brand of AI and constantly shifting definitions of AGI around 2025. When a bunch of startups started slinging around SI as a marketing term and of course when Meta launched MSI in summer of 2025.
that was all before Trump's "edict"
Why people downvote objective, non-controversial comments, is beyond me.
[dead]
how much benchmark-maxxing has been done here? taking bets on poly and kalshi...
I think it's safe to assume every single model is benchmaxxed. I mean why wouldn't you include training data from common benchmarks? Not only is a higher score good marketing, but these benchmarks are supposed to represent common tasks so training for those should in principle improve performance across a wider range of tasks.
There's an entire section in the linked post about measuring benchmaxxing, so I guess that probably says something.
Out of all the tech CEOs out there, Zuck would never lie to me or manipulate me. I know this in my heart.
The thing about Zuckerberg is that although he's deeply odd he has no ability to conceal this.
Sam Altman is much more dangerous because he is occasionally able to come across as human.
All of them are doing all the benchmaxing they can most likely. Dumb not to when so many people care about those numbers and don’t have the opportunity to try every model out personally.
Muse Spark was generally good. I'd be surprised if this model felt benchmaxxed.