Qwen3.8-27B

(twitter.com)

276 points | by mfiguiere 1 hour ago

39 comments

  • hypfer 15 minutes ago
    Since it might be helpful to some, here's my current commandline for llama.cpp running on an RTX 4090 with my monitor moved to the iGPU to free up all of its VRAM.

    llama-server -m Qwen3.8-27B-IQ4_NL.gguf --mmproj mmproj-BF16.gguf -c 170000 --parallel 1 -ngl -1 --cache-type-k q8_0 --cache-type-v q8_0 -b 1024 -ub 512 --flash-attn on --no-context-shift --no-mmproj-offload --spec-type draft-mtp --spec-draft-n-max 5 --spec-default --cache-type-k-draft q4_0 --cache-type-v-draft q4_0 --threads 24 --jinja --reasoning on -fit off

    Identical to the qwen3.6 config. With a prompt like "svg owl" (which can reuse quite a lot compared with creative writing or similar, so ngram-mod shines), I get about 70-80t/s like this, with a memory overclock of about 1.5GHz

    • reilly3000 4 minutes ago
      Thanks for posting! Have you had any success with running without kv cache quantization? Is there a noticeable difference in quality without any? I would assume that would eat into context but 170k is pretty generous!
      • hypfer 1 minute ago
        According to this shitty vibecoded thing "I" built https://hypfer.github.io/will-it-fit-llama-cpp/, FP16 K/V would give me something like 90k context at the same model quant, which doesn't really fit my usage.

        But maybe someone else has experience to share there

  • KronisLV 1 hour ago
    I hope really badly that we'll get a new 35B A3B or similar MoE model!

    I also miss the Qwen 3 Coder Next, which was 80B A3B, there are quite a few use cases where a non-dense model <100B would be the sweet spot (when you have the VRAM but not the TDP or compute power). Heck, I'd gladly take A5B or A8B or even A10B as a sort of middle ground.

    Also alternate link for viewing the images without signing in: https://xcancel.com/Alibaba_Qwen/status/2088280182356611304

    • jwr 5 minutes ago
      Me too. 35B A3B runs really fast on my MacBook Pro (M4 Max) and is suitable for real-time tasks like dictation post-processing. The dense model is not.
    • Casteil 1 hour ago
      I'm hoping too that they'll put out some MoE variants.

      Qwen3.5:122b:a10b can run about twice as fast as this 27b dense model.

      Edit: Like its predecessors, 3.8 seems really inclined to overthinking, and on a 27b dense model that's kind of painful. I think I'm going to stick with gemma4:26b-a3b as my go-to because it runs about 4x as fast and tends to only need a fraction of the tokens in its 'thinking' stage to get the same or similar answer.

    • peri-cl 1 hour ago
      Same here! Qwen3.6-35B-A3B is the only local model I've found that runs reasonably on my iGPU. Looks like me and and my noisily-wheezing laptop will be sitting out this upgrade.
      • expedited123 4 minutes ago
        Mind sharing your laptops specs? Just interested to see what is needed to locally run Qwen3.6-35B-A3B
    • Alifatisk 1 hour ago
      > I'd gladly take A5B or A8B or even A10B as a sort of middle ground.

      Whats up with focusing on the active param count? Do yall fiddle with the weights or something?

      • kennywinker 19 minutes ago
        Total param count decides how much vram you need to run it. Active param count decides how fast it runs. My 10 year old GPU can load quantized 35B or 27B, but it can’t process 27B parameters per token faster than 2-4tok/s, while it can do A3B at >40tok/s
      • martinald 1 hour ago
        You can run these on CPUs at a somewhat reasonable speed.
        • KronisLV 42 minutes ago
          Or (somewhat) low TDP GPUs for that matter, like workstation ones, that might have enough total VRAM but not the best bandwidth/compute.
  • scrlk 1 hour ago
    Beats Opus 4.7 Max (w/ Claude Code) on DeepSWE (42.2 vs 40). Looks like Qwen's 27B models continue to pack some punch.

    Unsloth's GGUF quants are up: https://huggingface.co/unsloth/Qwen3.8-27B-GGUF

    • NitpickLawyer 1 hour ago
      > Beats Opus 4.7 Max

      I'm a huge open model fan, and have used them since forever, even have daily drivers for on-prem dev, but no. They do not beat opus on real-world usage.

      Qwen models are impressively good for what they are, are "good enough" for plenty tasks, can be ran locally on decently priced hardware, and so on. They certainly have their uses, and the field in general has advanced faster than my early expectations. But to compare a 27B model to SotA behemoths from a few months ago is doing everyone a disservice, especially people who pick it up, try to use them just like API models, and leave disappointed and confused. Number goes up on a benchmark isn't it.

      • spmurrayzzz 1 hour ago
        > They do not beat opus on real-world usage

        We have an internal eval that measures performance on tasks for a handful of embedded systems repos for our mmWave radios (mostly Rust, some C for microcontroller stuff). Qwen3.6-27B scores only 4% lower for pass@1, n=250 compared to Opus-4.8.

        For the labeled dataset, the average PR size they're being measured against is around 1.5k SLOC.

        This is very much "real-world usage" for us. The sort of change sets that come in daily/weekly and are solving non-trivial issues in the respective codebases.

        As is usually the case, the most broad claims from both the labs and from the consequent pushback are talking past each other.

        • cyanydeez 11 minutes ago
          Let us know when you have Qwen vs Qwen comparison stats. As long as there's not a regression, that'd be awesome.
          • spmurrayzzz 8 minutes ago
            4% is within the margin of error anyways for pass@1, so I think pass@k > 1 is gonna be the better indicator of any movement (still need to calibrate the optimal k to re-test). 10 seems too tolerant even though that tends to be the next tranche I reach for.
      • KronisLV 1 hour ago
        > ...but no. They do not beat opus on real-world usage.

        I agree, but then we just need meaningful benchmarks that clearly show that! Otherwise it's hand waving about something that should be put on paper in quantifiable terms.

        • pimeys 46 minutes ago
          If you are working in a company and using language models, it is a very good idea to hold a bunch of evals you can trust and use to validate new models. Calibrate every once in a while with prod data. We have our own and the only numbers on quality and cost I trust come from this setup.
        • niek_pas 1 hour ago
          A wise man once said, "not everything that counts can be counted, and not everything that can be counted counts".
        • mlmonkey 28 minutes ago
          In the end, the only benchmark that matters is your own.
        • bewareofscams 1 hour ago
          Only useful benchmarks are those you (in particular) don't have access to.
        • xienze 1 hour ago
          > but then we just need meaningful benchmarks that clearly show that!

          That's the rub. AI benchmarks are IMO, by and large totally unreliable. We think of them as similar to traditional benchmarks of deterministic processes where the number of variables is low. But they're anything but that. Non-deterministic processes with an astounding number of variables and fuzzy acceptance criteria.

          It leads to results like these, where if you take it at face value, the only conclusion you can draw is "wow Anthropic must be stupid if Opus takes 1T parameters to do what Qwen can do in 27B."

      • metadat 14 minutes ago
        How can you say this when you haven't even tried it yet? Is it just hypothetical vibes?
      • willcmcc 19 minutes ago
        There is 0 shot you can make that claim about this model you have not used or downloaded yet
      • altmanaltman 9 minutes ago
        "Benchmark is stupid" and "model beats model on benchmark" are two different things, though. The second one is objectively true regardless of your views on the first one, right? To expect everyone to share your opinion that benchmarks are stupid is pretty weird, and just saying "no" to an objective truth is the definition of delusion.
    • nblgbg 1 hour ago
      Is there any advantage to using the model from Unsloth compared with https://huggingface.co/Qwen/Qwen3.8-27B-FP8 ?
      • benxh 1 hour ago
        Depends on what software/hardware you'll run it. GGUFs from Unsloth can run on pretty much every single potato; full weights need beefy gpus
      • danielhanchen 44 minutes ago
        We also made NVFP4 ones if that helps! https://huggingface.co/unsloth/Qwen3.8-27B-NVFP4
      • petu 1 hour ago
        Unsloth one is gguf for llama.cpp (and some other on-device engines).

        So advantage is not having to produce your own quantisation / gguf from .safetensors you've linked.

      • 4chandaily 1 hour ago
        Run the unsloth if you are using llama.cpp (GGUF)

        Run the one you linked if you are running vllm (safetensors)

    • Foobar8568 1 hour ago
      Considering the clusterfuck that is opus 5 or even fable, if Qwen 27B is trully better than Opus 4.7 Max, I will rejoice.
      • UncleOxidant 48 minutes ago
        If it's as good as Sonnet 4.6 for most things I'd be happy.
    • WithinReason 1 hour ago
      I wish each quant was benchmarked on the same tests as the original network so we could compare their performance
      • scrlk 59 minutes ago
        Unsloth publishes KL divergence numbers which measures how much the quantised probability distribution changes vs unquantised: https://unsloth.ai/docs/models/qwen3.8#quantization-analysis

        It's a bit bare at the moment, I assume they are going to add further detail later (eg comparison to other quants), similar to their other releases.

        • zargon 11 minutes ago
          KL divergence is nothing close to a replacement for benchmarks. As flawed as benchmarks are, KL divergence is a barely useful signal. The fact that Unsloth only just started publishing KL divergences shows how unserious the quantization space is.
    • edg5000 1 hour ago
      That's crazy, considering the massive size difference. But the small Qwen models are known for punching above their weight.
    • UncleOxidant 50 minutes ago
      Good morning Dario!
  • ramon156 1 hour ago
    People will claim it's not comparable to Opus despite it beating the score. I'm not sure I disagree, but I'm also unsure whether I care. Most new models nowadays are "good enough". I cannot complain because I'd rather spend that time improving my prompts and docs. Opus might be a _slight bit better_ at picking up vague hints, but it's also extremely expensive, and I hit the 5 hour limit way too quick.

    I care a lot about speed and efficiency right now. For my setup I would like to have 2-3 different model families. I've settled on GLM-5.3 (formerly Deepseek v4 pro 0813) for architecting, Deepseek V4 Pro 0813 for developing, and Gemini flash lite (any recent cheap model) for repo scouting. I'll add another one in the mix for reviewing (in this case Gemini 3.7) and that's all I need.

    I've tried most models except Grok.

    Qwen is too expensive IMO (Alibaba Cloud subscriptions are hard to come by and I'm not spending 50 euros a month for a tool, so 18 euros it is). If it ever becomes efficient enough to run locally I will definitely look back.

    Claude is slow and expensive (the cache hit prices are absurd).

    OAI is pretty good, I might add it to my arsenal seeing how cheap it is.

    These opinions change every day. Last week I would've never picked Deepseek until I read about the pricing. even post aug 16 it's worth it (although it's getting close to gemini pricing).

    Right now my costs are 12 euros a month (z.ai) + whatever deepseek consumes. This typically isn't more than 8 euros a week. 44 euros a month and I have a setup that is doing pretty well.

    • simplyluke 41 minutes ago
      I'm convinced a lot of the anti-open-weight model comments at this point are inorganic traffic - there's trillions in investor money riding on a world where these models aren't cheap commodities. Having actually used things like the recent GLM, Kimi, and Qwen I think any edge the labs have is marginal at most and actually prefer the open weight models in most day to day usage.

      Anthropic's recent releases are wordy to the point of exhaustion. Every time I use opus recently I find myself wanting to yell "GET TO THE POINT" at a terminal, which is exacerbated by it being slow.

    • hypfer 1 hour ago
      > I've settled on GLM-5.3 (formerly Deepseek v4 pro 0813) for architecting

      Dude, GLM-5.3 released _today_.

      The phrasing "I've settled on" is incorrect for this context.

      • ramon156 1 hour ago
        hence the "former deepseek v4 pro". I tried it out this morning and have had no complaints. I already liked glm 5.2
        • Topfi 6 minutes ago
          Honest question, how do you assess models this quickly? What metrics are you using? Would love to get my suite from multiple days and hundreds of prompts down to minutes. Got a few first pass tasks I run upon release for an initial experience, but those only work because even Fable and Sol fail despite objectively correct solutions existing, so it works because most models fail, but then, those are consciously not enough for coding, tool use, adherence or task specific inference and assessment…
        • hypfer 59 minutes ago
          The sentence still doesn't make sense, because "settled on" implies a long testing phase with a verdict eventually emerging out of that.

          What you're currently doing is "testing out"

    • tosh 37 minutes ago
      i think you will like luna if you haven't tried it yet
  • Casteil 43 minutes ago
    One thing a lot of people don't seem to factor when hyping Qwen is how much models like this tend to 'overthink' with seemingly endless 'second guessing'. 3.8 seems no different from what I've tried thus far.

    As capable as it is, it's hard to justify using it when a competing model (e.g. Gemma4:26b-a3b) can consistently achieve the same or similar response with only 1/10th as many 'thinking' tokens, achieve much higher tokens/second, and take a small fraction of the time. I suppose 'YMMV' depending on your use case.

    Also, I haven't used it enough yet to see if it's prone to infinite looping, but its predecessors sure were.

    • lrvick 39 minutes ago
      Use 3.6 27b as a daily driver for months with charmbracelet crush. Gemma 26b-A3b is not even remotely comparable in terms of coding for me. YMMV depending on how you work, what harness you use, etc I suppose.
    • ThouYS 34 minutes ago
      gemma4 can't hold a candle to 3.6
    • cyanydeez 31 minutes ago
      You can add a thinking budget thats not much effort in llamacpp. You can align the cut off message with your agent instructions.

      What you describe is a engineering harness problem.

      If you, and i mean the royal you, actually read tge thinking traces you can see and figure out where its stuck

      This means an effective harness would observe when the model is overthinking and step in with reasonable redirection, like increasing logging.

      Llamacpp can set reasoning budget and message per reauest, so it can be dynamic.

      Your complaint is "skill issue" based and will be resolved by people who do something ither than vibe code react demos.

  • LeBit 1 hour ago
  • mraza007 14 minutes ago
    Man what a week, We just had GLM 5.3 that came out and then we had smaller local model Qwen3.8-27B from Qwen

    Just tried using Pi Agent and looks very promising

  • T0mSIlver 1 hour ago
    Unsloth Q4_K_M on a single 3090, llama.cpp "Generate an SVG of a pelican riding a bicycle" first try https://www.reddit.com/r/LocalLLaMA/comments/1voa3ch/comment...
  • TomGarden 1 hour ago
    Any tips on the best approach at running this at an M4 Max 128GB? Token throughput was a bit slow with the last 27B one (MLX), ended up using the A3B variant but if I could get this one to reasonable speed I'd much prefer it.
    • seanmcdirmid 6 minutes ago
      27B is a dense model so it will be slower with an MoE (A3B), but should have better quality? I still haven’t found very good uses cases on my M3 Max for dense models. Even if you can find a MTP version, it doesn’t help much, especially if you compare against an MoE with MTP as well.
    • mft_ 1 hour ago
      Go for a slightly more quantised version, and experiment with different MTP settings. I find that MLX versions are marginally faster on my 64GB M1 Max, but I usually use Unsloth's GGUFs via llama.cpp as there's a much greater range of quants available and I prefer llama.cpp. MTP sometimes also helps a little, but I suspect it's less helpful on my system than others.

      Unsloth: https://huggingface.co/unsloth/Qwen3.8-27B-GGUF

      This might work for you, but I didn't get on very well with MTPLX when I tried it a while back; YMMV: https://huggingface.co/Youssofal/Qwen3.8-27B-MTPLX-Optimized...

      • evgen 53 minutes ago
        This is the way if you need speed. It costs a little bit in smarts, but compare the MTPLX option listed above with the oQ4e-mtp quant using oMLX. The good cacheing layer in oMLX will help things feel faster for some classes of tasks in my experience.
    • UncleOxidant 41 minutes ago
      Wait for the MTP variants that will likely be out within days. I'm on a 128GB Strix Halo box and for 3.6-27B 8bits I was getting about 9tok/sec (not great). With MTP that gets closer to 18 tok/sec (kind'a usable).
      • anana_ 27 minutes ago
        Seems like MTP is available immediately!
    • LoganDark 1 hour ago
      Unfortunately, that chip just doesn't really have the memory bandwidth to run this (or nearly any) model at acceptable speeds. I have the exact same chip (M4 Max 128GB) and I've been trying to optimize a completely purpose-built implementation with Fable and this is just not possible. Even if you could reach the full 576GB/s, it's just physically impossible to exceed these numbers with the model's architecture:

      2 bpw - ~85.7t/s

      3 bpw - ~58.0t/s

      4 bpw - ~43.9t/s

      6 bpw - ~29.5t/s

      8 bpw - ~22.2t/s

      16 bpw - ~11.2t/s

      without cheating. You'd have to exclude layers, skip operations, etc. basically do stuff the model wasn't trained for. And speed collapses so fast with context that even 2 bpw would be looking at ~37.6t/s after just 128K tokens.

      MTP only improves the situation by up to 2x in the ideal case, while drastically reducing the performance floor. While optimizing a 9B model on this hardware, I've found that the GPU just doesn't have enough FLOPS to handle speculating more than one or two tokens ahead on a single stream, regardless of quant level, simply because of the arithmetic cost of the forward pass. The 27B model would be even more expensive than that, potentially such that it's already bottlenecked by the GPU itself rather than memory.

      I wouldn't get my hopes up for the 35B-A3B either. Not only is it reportedly much less intelligent, but I hit the exact same 85t/s wall in practice (again with highly specialized inference).

      Without speculation I can reach around 120t/s on Qwen3.5-9B and with n-gram speculation (not even MTP; this derivative didn't come with one) around about 150t/s on average. This is on the very very edge of what I'd consider acceptable for me to even consider using such a compact model. YMMV due to the silicon lottery but the situation isn't good.

    • brcmthrowaway 1 hour ago
      Check out MTPLX and limit your context size.
  • jedbrooke 1 hour ago
    I hope the bonsai team makes another 1bit quant of this model (or releases code/instructions on how to do it), using the Qwen3.6 27B on my 16GB mac mini has been wild . The 1bit quant feels like opus level… for the first couple turns. Then it has trouble eg switching from plan mode to act mode. This is mostly mitigated by starting a new session. (tbf this limitation is called out on the hf page)

    I saw unsloth has 1bit quants too so I might check that out, anybody have experience with those?

    • spwa4 23 minutes ago
      Sounds like you need to check what the max context is set to ...
      • jedbrooke 7 minutes ago
        100k is all the context I have ram for, this is with any auto-compact turned off. This is using Cline in vs code. I’m sure I could tune the system prompt and mode switching more to work better with this specific model, but I haven’t gone down the custom harness rabbit hole yet.

        And this is also specifically for the 1bit quant version. I don’t think the fp8 or even fp4 versions have this issue, but I haven’t tried those much

  • NorwegianDude 1 hour ago
    If the benchmarks are a real indication, we now have a local model that is runnable on a high-end personal PC that trades blows with the leading model Claude Opus 4.6 Max from half a year ago.

    Insane if that is the case. Downloading now!

  • cmrdporcupine 1 minute ago
    I found this kind of amusing while running it (using Pi as the harness). Don't know if this is evidence of intense fine tuning from Claude but it smells like it...

    " The user wants me to explore the repository at XXXX and report back. Let me start by understanding the project structure, reading the CLAUDE.md file, and getting a general overview of what this repository is.

    Let me start by reading the main project documentation and exploring the directory structure.

    I'll take a look around this repo. Let me start by getting a lay of the land.

    read resource CLAUDE.md (ctrl+o to expand)

    ENOENT: no such file or directory, access 'XXXX/CLAUDE.md'"

  • tosh 1 hour ago
    27b dense model at Opus 4.6 level

    Opus at home

    I hope there also will be a new ~10b variant

    • UncleOxidant 39 minutes ago
      I'm hoping for a 3.8-122B MoE
    • yassa9 1 hour ago
      can you tell me ideas of usecases of 9 or 10B language models ? I cant find any usecases other than training a lora on them to give good bash commands for example
      • mring33621 38 minutes ago
        9B Qwen models are good and fast for local python coding tasks.
      • tosh 58 minutes ago
        they are all overlapping but:

        categorization, information retrieval, semantic search, image description

        also with the model as part of an agentic system with tool calling

        (edit: it is quite impressive what a small model in a feedback loop can do)

  • minimaltom 37 minutes ago
    Architecture thread! Afaict they continue to use gated attention + delta net, which was also adopted+adapted by K3, but im surprised theres no improvements to the residual stream (deepseek are using manifold hyper-connections, kimi have attention residuals) ?

    Perf improvements seem to all come from training?

    • anana_ 29 minutes ago
      As was the case with GLM 5.3, it seems that there is still much juice to be squeezed from post-training
  • chvid 1 hour ago
    These are massive improvements - and something you can actually run on a laptop.
  • arjie 31 minutes ago
    I use the Qwens as a vision model for my DeepSeek V4 Flashes to handle. But the Qwens run on old RTX A6000 Ampere. Does anyone know if there's any news about INT4/AWQ quants for the RTX A6000?
    • ericd 21 minutes ago
      Was recently thinking about doing something similar, do you basically just have the qwens describe what they see for the flashes?

      Was considering adding a LoRa/vision head to Flash, but seems like it could take a while to get it right.

      If DSv4 Flash was multimodal, I’d probably be done model shopping for a while

  • mickeyp 1 hour ago
    Model benchmarks are useful, to a point, but it is the long tail of things you do with the model that determines if it's good at a wide range of activities. Ant/OAI, to their credit, build their models -- even the small ones -- so they follow instructions and do tool calling well, without the system prompts confusing them. This is especially important for long-horizon tool calling.

    So one open weight model might "meet" Opus or whatever on benchmarks, but then fail to follow a simple answer format and also tool call correctly. The models are whipped to within an inch of their lives to strictly adhere to their post training quality gates.

  • theanonymousone 1 hour ago
    I'm wondering whether any provider can offer this for cheaper $/token than the new DSv4 Flash, which is both cheaper and smarter :/

    Completely local use is a different story, of course.

  • irthomasthomas 33 minutes ago
    Why don't qwen/alibaba host the model themselves? I was looking forward to trying it on their coding plan. Google are the same way with their Gemma models.
    • spwa4 22 minutes ago
      Pretty sure you can use Gemma models on Google's "Vertex AI".
  • jlkivey 57 minutes ago
    Note: on the model card the comparison to Opus is Opus 4.6 Max, not 4.7
  • bertili 27 minutes ago
    Wow. Speed improved as well. 200t/s on a RTX 5090!

    https://x.com/sgl_project/status/2088281320422322413

  • yassa9 1 hour ago
    Can anyone who has that specific personal test he tries on different models , and tries this model , to tell us here if possible , how good or bad is this new model ? compared to others ?

    I only trust those users genuine personal tests

    • alyandon 1 hour ago
      There is a down to earth guy on YT that performs a series of tests against LLMs running on non-god-tier commodity hardware. He will likely be testing this soon enough.

      https://www.youtube.com/@lukesdevlab

      I don't know if that is what you are looking for or not and as always your experiences may be different.

      • yassa9 33 minutes ago
        thaaanks man, this channel seems really informative, although < 10K subs only !
        • alyandon 27 minutes ago
          It's a relatively new channel - but yeah - I feel the guy puts a lot of effort into what he does and deserves more subs.
  • kunver 1 hour ago
    Looks like a pretty significant improvement on the DeepSWE benchmark compared to the previous 27B model.
  • ThouYS 1 hour ago
    3.6-27B on little-coder was already mind blowing. looking forward to this guy!
  • kristopolous 1 hour ago
    q4km is about 48 tps on a 4090. my llama.cpp params are --flash-attn on --parallel 1 --load-mode mmap
  • hathym 1 hour ago
    better than opus 4.6 max ╰(°□°)╯

      __        __   ___   __        __
      \ \      / /  / _ \  \ \      / /
       \ \ /\ / /  | | | |  \ \ /\ / / 
        \ V  V /   | |_| |   \ V  V /  
         \_/\_/     \___/     \_/\_/
  • anana_ 1 hour ago
    Monstrous benchmarks! Hoping it is not benchmaxxed.
  • tosh 1 hour ago
    also cool: Qwen 3.8 27b is multi modal!
    • gurkwart 1 hour ago
      strong visual reasoning apparently, which is nice. still lacking native audio however. hoping for more companies to embrace the spirit of something like `gemma-4-12b-qat` for actual multi-modality (text, image, video, audio).
  • pu_pe 1 hour ago
    Seems to be SOTA for its size. Hopefully independent benchmarks will come soon.
  • kunver 1 hour ago
    Welcome deepseek flash flash!
  • alpha_trion 1 hour ago
    NICE, i've been waiting for this drop, thanks for posting this
  • filup 1 hour ago
    https://news.ycombinator.com/item?id=48403639

    my prediction was way too far out. 4.6 at home! Woo.

  • altruios 1 hour ago
    remember to let llama.cpp catch up to anything new in this model. Save your judgment until about 2 weeks of use.
    • chrismartin 1 hour ago
      'Good' news, there seems to be nothing new architecture-wise. Same as Qwen 3.5 and 3.6, so llama.cpp doesn't know the difference.
  • brcmthrowaway 1 hour ago
    This with ddg mcp to fill in world knowledge. Are local models the future when computer architectures catch up?
  • WithinReason 1 hour ago
  • brcmthrowaway 1 hour ago
    My Strix Halo is about to go overdrive!
  • ramon156 1 hour ago
    need another fable uncensored merge with 3.8, really curious what it can deliver
  • RobertasTa 23 minutes ago
    [flagged]
  • fintuner 35 minutes ago
    [flagged]