Meta Muse Glimmer – open weights 30B local coding model

(research.meta.ai)

245 points | by riordan 2 hours ago ago

92 comments

  • scrlk 2 hours ago

    Will be interesting to see how Qwen3.8 27B compares against this once it releases this week. Seems like dense 30B is back in fashion?

    EDIT: An open weight version of Muse Spark 1.2 is going to be released as well:

    https://x.com/alexandr_wang/status/2086756152034066792

    https://xcancel.com/alexandr_wang/status/2086756152034066792

    • pu_pe an hour ago

      Based on the benchmarks, it seems that Muse Glimmer barely edges out against Qwen3.6 27B, except for tool-calling skills (MCP, etc.). I wouldn't be surprised if they released it now because they are afraid they wouldn't beat Qwen3.8 27B.

      • mycall 41 minutes ago

        Do AI companies make release plans based on upcoming other models like this? I would think all the processes that go into the repository and weight infrastructure pre-training, checkpointing, knowledge distillation, model compression, post training pipeline, ecosystem integrations, inference API, benchmarking, human eval/safety/alignment, docs, etc... all that dictates the release schedule.

        • michimagdesign 20 minutes ago

          Yes, not every model release is reactionary to other labs. Either they had hints for the release of other models or they cut efforts in late stage testing of the models to hit these earlier release dates. There’s always some flexibility. And there’s certainly the incentive to cannibalize the news cycles for competitor models.

        • stogot 18 minutes ago

          the last few items there (benchmarking, human evaluation, docs) can be rushed or skipped by leadership if they want to beat comp. they probably spend a few weeks on those things normally

        • echelon 29 minutes ago

          There has been a long history of AI model releases made shortly before or after a major planned release by another company. Almost always to upstage or steal thunder.

          Just recently, Minimax H3 released as open weights on the eve of Seedance 2.5 global availability. It's not as good, but it's good enough and it's completely open.

          Flux 3, which is nowhere near as good as either, suddenly announced their release once news of these other two became public. They knew if they waited they'd be ignored. It didn't really help them much, unfortunately.

          The LLM releases are even more rivalrous.

          And don't forget all of the competing launches planned before Google IO or major release events.

          Companies like to eat into the news and press cycle of their rivals.

          • Sabinus 19 minutes ago

            I've seen it here on HN (it's particularly noticeable via the /active page) multiple times. If Google, OpenAI or Anthropic release something significant, odds are good you'll see a headline from one of the others.

          • Forgeties79 3 minutes ago

            >long history

            Seems a bit premature of a statement lol

    • karimf an hour ago

      Yes, and also waiting for the next iteration of Gemma. Muse or Qwen are optimized for coding, while IMO Gemma is still better for non-coding tasks.

      https://x.com/osanseviero/status/2086107547535122767

    • Gecko4072 an hour ago

      Makes me feel hopeful. Things felt more positive around the llama 3 era. Now it’s like a dark, dreadful race.

    • wronglebowski 2 hours ago

      It’s really interesting timing, Qwen over thinking is what kills it for me. I’m just glad we have more options in this size class now.

    • imilev 29 minutes ago

      yes i think everyone is waiting to see that ;d, i've been on qwen for the last year and a half now.

    • lostmsu an hour ago

      It seems worse than 3.6, but a bit smaller.

      • IsTom an hour ago

        How is 30B smaller than 27B?

        • lostmsu 3 minutes ago

          [delayed]

        • LeBit 43 minutes ago

          It uses fractal compression

    • ignoramous 42 minutes ago

      > Seems like dense 30B is back in fashion?

      Surprising that Meta don't host this model, even as rate-limited free-tier.

      > open weight version of Muse Spark 1.2

      Wait. Is this "version" different from what Meta serves?

  • sajithdilshan 2 hours ago

    Still needs 32-64GB memory to run it locally. 64GB Macbook pro with an M5 chip costs more than 4k Euros in Germany. A more practical model would be a language specific (e.g Python or JVM language) and excellent at tool calling and reasoning. Maybe that way they can shrink it even more.

    • karimf an hour ago

      Practically ~20GB with KV cache

      > We quantize weights to ~4-bit, bringing the LM under 20 GB. We validated minimal to no degradation on agentic tasks under compression.

      https://www.reddit.com/r/LocalLLaMA/comments/1vkgsum/introdu...

    • mihaelm an hour ago

      I'm sooo happy I pulled the trigger on upgrading and getting a new laptop (with 64 GB RAM) last summer. Feels like it was just in time before the exponential price jumps.

      • ishtanbul an hour ago

        Pulled the trigger?

        • mihaelm an hour ago

          lol, you're right, the brainfart completely changes the meaning.

          I corrected it.

        • idiotsecant an hour ago

          Common phrase.

          • karolist an hour ago

            Parent used "pulled the plug", are you saying it's applicable here and not "pulled the trigger" like suggested?

          • Hinrik an hour ago

            That commenter you're replying to knows that. The original commenter before them wrote "pulled the plug" which is different and doesn't quite apply here (actually implies the opposite of what they meant to say).

      • mettamage an hour ago

        Bought an M1 64 GB for 2000 euro’s second hand a year ago. That was sweet

        • karolist an hour ago

          paid 2.7k € for this same build new in Dec 2023, that was also sweet (still is)

    • dbbk an hour ago

      Well if you're spending thousands on API tokens already, you could just drop the same amount on a 128GB MacBook Pro and that's a one time cost.

      • smallerize 29 minutes ago

        If you're dropping thousands on API tokens, you're going to be slowed down at least 10x trying to do everything on a single MBP.

      • neuroticnews25 31 minutes ago

        Don't forget about energy usage, you'll probably never break even vs same model on openrouter.

        • jurgenburgen 18 minutes ago

          If you can’t do it cheaper on your own hardware it does make you wonder how much of the cost of inference those large LLM providers are eating? Datacenter hardware isn’t magic.

    • Gecko4072 an hour ago

      There have been discussions on language specific not really being a relevant change to reduce size.

      • Manfrednotfunny an hour ago

        I would love to see any good research projects about it but i have the feeling that Frontier with MoE is making too fast of a progress so that a customized model would always be worse and that the MoE part is actually going somehow in this direction.

        On the other hand, at the GTC was a talk about coding in different lanugage (like spanish) and explaining that the quality between spanish and english is relevant different.

        But i have not found a good article about the impact of learning data with practical experiments or even if the order of the learning data matters.

        At least I think i remember that Meta mentioned having better and less data can be better than more data with lower quality.

        As long as these models can explain to you facts about any other topics, its still overfitted for the task though.

        • mapontosevenths 27 minutes ago

          Capability in LLM's is distributed throughout the manifold in subspaces. Even worse, the subspaces exist in superposition.

          That is to say, there is no single 'python' part of the model. The python bit is spread throughout the entire model and overlaps with other pieces that have similar, but unrelated, capabilities. For example the python subpspace might be partially in superposition with cupcake recipes, Esperanto, and calculus. We need calculus in a coding agent but not the other two. However, separating them cleanly is almost impossible, and even identifying them is tough.

          Internally the manifolds are highly inefficient and nothing like you would imagine something humans built would be designed. It's more like something that evolved in nature.

          • Manfrednotfunny 23 minutes ago

            My current image from a MoE is that the base/core might be the more generic thing and that things like python are part of one expert though.

          • dist-epoch 6 minutes ago

            There was some paper about routing at training bio-knowledge into a particular region of the model, which you then can cutoff when serving. But you probably lose some efficiency since maybe you sized that region too small/too big.

    • solarkraft an hour ago

      I feel like we’ve had this discussion before. From what I remember, specialized models rarely do that much better than general ones, hence no mode Codex models.

    • formerly_proven an hour ago

      4K bucks buys you around 180 months of <insert AI subscription here> with zero upfront cost.

      • zamalek 26 minutes ago

        Problem is that might go away or get nerfed.

      • Mistletoe 39 minutes ago

        Haha wow. I’m trying to even imagine the AI landscape in 15 years and I can’t.

    • cynicalsecurity 37 minutes ago

      I don't understand the desire to run own AI models for programming locally. No laptop is ever going to be as powerful and energy efficient to run anything close to OpenAI, Anthropic or Google models. A model you can run on a loptop is simply not going to work as well as it's needed for programming. Small models for linguistic work fine, but anything more sophisticated simply won't provide enough resources or power. Or models would need to be significantly dumbed down - then why use them at all? So far the idea of carrying a "thin" or "thin"-like device looks more reasonable to me, while running AI on your own server.

      • OtherShrezzing 25 minutes ago

        > A model you can run on a loptop is simply not going to work as well as it's needed for programming

        The models you can run on a high-spec laptop today are approximately where frontier models were 12-18mo ago (albeit at a lower tok/s rate). If you scan back through hn comments from that era, you’ll find plenty of people saying “this is powerful enough to massively increase my productivity”.

      • ComputerPerson 24 minutes ago

        I've never done it but would be interested because it cuts out the burden of worrying about costs. Maybe I'm mistaken on energy cost here. There's a constant raincloud that follows me around regarding limits, and it would be nice to shake that.

        I've been able to accomplish incredible feats (for myself) since GPT-4, so model intelligence is secondary.

    • sparkling an hour ago

      Even if you had a 64GB machine: Are you willing to reserve 90% of your memory to run a LLM? With dirt cheap models like deepseek-v4-flash that will run "forever" on $10, the answer for me is clearly: no.

      • Manfrednotfunny an hour ago

        I'm waiting for the speed/quality per dollar metric to go down a little bit further and then I will def run it at home.

        Its not just that you send a sentence to an API endpoint, you always send EVERYTHING to that agent as a context.

        You want to analyse your spending history? You now send everything to someone.

        Either no one cares but understands this implication on how easy it is to really capture you or no one really things about it.

        But i'm a lot more diligent on what I send. I disabled the gemini activity feature for example because google started telling me that my stuff could be reviwed by humans.

        • plufz 33 minutes ago

          Yeah, it does feel a bit silly with my encrypted disks, encrypted backups, unique passwords, advanced router, etc, while I send everything I do in plain text to anthropic.

      • zoobab an hour ago

        "With dirt cheap models like deepseek-v4-flash that will run "forever" on $10, the answer for me is clearly: no."

        When it's free, you are the product.

        • amrit3128 15 minutes ago

          Ah yes, I'm sure Trovalds and Stallman are harvesting my data through free software, aren't they? This argument is used by boomers who were fed cold war era propoganda that surely everybody is selfish, and you're always at fault.

        • prplxd_nihilist an hour ago

          I see many people saying deepseek and other chinese providers have always been profitable. Also they show their training costs publicly. Can't say for sure since I have not used it personally, but I think they'll for sure outlive the western SOTAs.

        • IMTDb an hour ago

          Deepseek flash is open weight, this means we can download and run that model without any connection to deepseek, no data/tokens/usage data ever reaches them. They cannot make us their product.

      • halJordan an hour ago

        It's the size of a big vm. There's nothing wrong with reserving that much working space for one item.

  • polymorph1sm 19 minutes ago

    Some interesting findings from the chat template designs:

    1. The template name is Onyx ATEM as found in the tool call exception message

    2. It appears to be following a harmony-style chat template. But the tool use seems to be a xml like :<atem:function_calls> / <atem:invoke> / <atem:parameter>

    3. atem: a internal joke of meta in reverse?

    https://huggingface.co/meta-models/Muse-Glimmer-30B/blob/mai...

    • dannyw 3 minutes ago

      The XML tags are similar to <antml:xxx>, which is obviously Anthropic ML (or ANTrophic xML).

      I think it’s likely 3; meta in reverse. While tokenisers and preprocessing can catch it, you want your special tokens to be unique and not present in the original corpus. <meta: is likely too common.

  • Gecko4072 2 hours ago

    What I think would be perfect is a model that could run on a single DGX spark and be competitive with DSV4 Flash 731. Flash is already a game changer. Hopefully meta plans on this, like the old 70b. V4 flash is smart enough for any use but slightly too big. 27b-30b isn’t intelligent enough.

    • 127 an hour ago

      DSV4 Flash 0731 already runs on RTX 4090 24GB + 128GB system RAM at a usable tok/s and quantization.

      • Gecko4072 an hour ago

        You personally? Just curious. Context window is also a factor and ram isn’t really cheap. Sparks are assembled units which I like.

    • cmrdporcupine an hour ago

      This model I think will be too slow for that on Spark, even at 4 bit quant.

      It's a dense model, not MoE like e.g. Qwen 35b or Gemma 4 26B A4B. On a Spark it will be memory bandwidth limited

      I haven't tried yet (working on it) but back of the napkin estimate puts it at around 15tok/s even after converting to NVFP4. Prefill would be much higher though. That 15tok/sec is pretty typical for dense models of this size:

      NVFP4 Q/K/V/O and MLP projections: ~13 GB/token

      BF16 attention gates: ~3 GB/token

      BF16 LM head: ~2.5 GB/token

      Total: ~18.9 GB/token

      At 273 GB/s, that gives a bandwidth-only ceiling of about 14.5 tok/s; actual performance would be lower.

  • _ache_ an hour ago

    It is interesting but it does look like a careful distillation of (Spark and) biggers open-weight models.

    The progress compared to Qwen3.6 27B is good, not that impressive, it's a 4 months old model. (kuto to them to compare to 27B dense and not 35B MoE, it's more fair to do so). It is very probable that Qwen3.8 27B will crush Glimmer-30B on most benchmarks.

  • vibe42 24 minutes ago

    Meta released their own 4-bit quant of this model for devices with 24GB VRAM.

    That's a modern gaming laptop; cheapest I see in the US with 24GB is $3.5k.

    Should be quite a bit faster than the new M5 MacBook Pro, and you can run Linux on it!

  • gunalx 2 hours ago

    Meta did not abandon opensource. I would love to see a smaller distill, or a moe of this size but the benchmarks seems competetive as long as it isnt benchmaxed witch i would not be suprosed if it is.

    • ignoramous an hour ago

      > Meta did not abandon opensource

      Open weights*

      I don't think outside of the Big 3 (Ant, OAI, GDM), given the strong competition from China, any other Lab has a chance at capturing the coding market if they aren't open weights (save for xAI whose latest Grok looks every bit good & will probably rely on Cursor for distribution instead of going open weights). There's literally no other selling point, as the capabilities have mostly converged by now among the chasing pack.

      • ComputerPerson 17 minutes ago

        There was a good discussion yesterday on the DeepSeek Flash release thread about this.

        There's a large market, very large, who want the best regardless of what it costs. Probably a large enough market to keep that domain of research afloat (as opposed to shifting research manpower to cost cutting).

        The reasoning is just that the marginal cost of AI is very secondary to fixed costs of the businesses themselves; it's not an excuse to sacrifice performance.

  • maxignol an hour ago

    Optimizing speed is really the way to go. Yet 24GB is not what everyone can afford. Maybe we could take some of those 56tk/s and transfer into some free RAM space using MoE loading ? I'd be glad with a less than 10GB and more than 6tk/s model.

    • lisplist an hour ago

      Unfortunately this is just the entry price for LLMs. With the exception of the Qwen 27B models, I personally haven’t found a ton of use cases for models less than 200B. With the right setup, fine tuning, etc, you can make small models do cool things, but hard to please everyone given the insane hardware costs at the moment and the comparably cheap API costs.

      • dist-epoch 3 minutes ago

        Gemma4-E4B (4B params) works pretty well as a local wiki, or when you don't have connectivity.

    • Manfrednotfunny an hour ago

      I don't thinnk just MoE will solve it. If you hit constantly different expert layers, you can't outsource layers efficently and have to swap it in.

      MoE will be faster because it will read less memory for sure, you still have to have it though.

  • tosh 2 hours ago

    good to see new open weights releases from meta

    • jauntywundrkind 2 hours ago

      good looking showing too, which is excellent.

    • InfiniteLoup 2 hours ago

      The least they could do, after ruthlessly bombarding my employer's servers with requests, ignoring the robots.txt, scraping everything, and incurring significant Google Maps costs for us in the process.

  • bentt 41 minutes ago

    Meta seems like the one American bigtech that would distill the the other American frontier models. My enemy’s enemy is my friend?

    • grim_io 39 minutes ago

      They do distill, their own bigger Muse model.

  • zmmmmm an hour ago

    Meta knows how to win back developer's hearts .... let's see if they have the goods

    • xandrius an hour ago

      If there is anything meta can do to regain hearts other than owning up their evil deeds, radically change their business model and paying up for taxes and damages, then the world is truly fucked and corporations will continue to win.

  • richardfey 2 hours ago

    Looking forward to giving this a try with llama.cpp. I’m watching the open-weights competition with high expectations.

  • bronxbomber92 an hour ago

    I wish they would release the quantized versions in a safetensor format. Many frameworks can't load PTE and GGUF.

  • sgt 23 minutes ago

    Can I run this on my RTX 5090?

  • Havoc an hour ago

    The favourable comparisons to Gemma 4 and qwen3.6 look promising!

    • cmrdporcupine an hour ago

      Those two offer MoE variants, this doesn't seem to.

      Dense model makes it dog slow on anything without HBM. Max 15tok/sec on decode on DDR5 systems like a Spark or a Strix Halo -- and that's at 4 bit quant.

      • EddieRingle 3 minutes ago

        Dense models run at a very usable speed (Qwen 3.6 was running at ~50t/s last I looked) on my dual 7900 XTX desktop. (And before anyone brings it up, I did not buy them for this purpose, so the up-front cost is irrelevant in my case.)

      • petu an hour ago

        3090/4090 probably would do 40 t/s, for 5090 75 t/s is shown in the blog.

  • solarkraft 2 hours ago

    Wow, Meta is back (at least for now)!

    I like this class of model. Multi-token prediction makes it viable to run dense models at not-too-far-off speeds as MoE models with much better intelligence.

    The submission’s title (open weights 30B local coding model) is luckily wrong: This is meant to be a general agentic model.

    It even comes pre-quantized and with a MTP/drafter model. Looking good!

    Let’s hope they aren’t dishonest with the benchmarks this time …

  • nutjob2 an hour ago

    The more open weight models get released the greater the market for personal and small business oriented hardware to run these models. This will drive lower cost hardware, which has stagnated in recent years due to most software not needing the performance and capacity.

    • grim_io 20 minutes ago

      Higher demand for 5090's did not make them cheaper, because Nvidia got much higher margin products to focus on.

    • cmrdporcupine an hour ago

      The opposite happening because foundries are full to capacity making higher margin stuff.

  • petcat an hour ago

    As an industry, I wish we would stop calling these things "open weight" because it is too easy to confuse with actual "open source", which they are not.

    Photoshop source code+ OSI license = open source

    Photoshop binary you can run on your own computer = open weight

    Photoshop SaaS web app = closed, proprietary (Opus, GPT, etc.)

    "Open weight" models are still just binary blobs that are completely inscrutable. It's like bringing home a dog from the rescue and just hoping that it doesn't have a tendency to bite kids in the face. You just can't know. The only thing that you can do is try to add more training (fine tuning) telling it not to bite kids.

    I don't think the FOSS community has ever accepted this, but somehow we're feeling like it is okay now.

    • microtonal an hour ago

      Photoshop source code+ OSI license = open source

      Photoshop binary you can run on your own computer = open weight

      I don't think this is a correct analogy. You are not allowed to distribute modified versions of the Photoshop binary. Most open weight model licenses allow you to make and distribute your own finetunes, etc.

    • craigmart 39 minutes ago

      I believe that comparing LLMs with traditional deterministic software is fundamentally misleading. It is extremely difficult to truly interpret what LLMs do internally, and as of now, nobody fully understands it. Even if you trained the LLM yourself, there is no source code you can simply read and learn from.

      Sure, having information about how these models were trained is helpful for reproducibility, but it is basically impossible for anyone without substantial capital and access to the same (likely copyrighted) data to reproduce the model. For normal users, owning the model weights essentially means owning 100% of the model, you can inspect and study the weights in much the same way as the lab that produced the model can, you can modify the weights, and you can use and distribute them if the license allows you to

    • QuadmasterXLII an hour ago

      Given an open weights model trained to sometimes bite kids, we can’t train it to not bite kids, even though billions of dollars of research have been thrown at this open problem.

      Given an open weights model trained to never bite kids, you can get it to bite kids with 10 prompts and a linear projection, the known simple algorithm doesn’t even need a backwards pass.

      yay asymmetry!

      • kzrdude 43 minutes ago

        Any pointers to more info about that? Sounds interesting.

    • piker an hour ago

      It is useful to indicate you can run the weights on your own hardware. That’s categorically different from most other commercial offerings. It’s as if your adobe example ignores the reality that would exist had photoshop been invented in 2019: cloud only.

    • monster_truck an hour ago

      This analogy is terrible and seems to be extremely misinformed about how rescues evaluate dogs before they are put up for adoption

      • petcat an hour ago

        I am extremely well aware of how rescues evaluate dogs. And I'm also fully aware that they do not know the full history of the dog. They go through a limited set of testing and interrogation to evaluate the safety of the dog. That's it.