Gemma 3 QAT Models: Bringing AI to Consumer GPUs

(developers.googleblog.com)

329 points | by emrah 8 hours ago

28 comments

  • simonw 6 hours ago
    I think gemma-3-27b-it-qat-4bit is my new favorite local model - or at least it's right up there with Mistral Small 3.1 24B.

    I've been trying it on an M2 64GB via both Ollama and MLX. It's very, very good, and it only uses ~22Gb (via Ollama) or ~15GB (MLX) leaving plenty of memory for running other apps.

    Some notes here: https://simonwillison.net/2025/Apr/19/gemma-3-qat-models/

    Last night I had it write me a complete plugin for my LLM tool like this:

      llm install llm-mlx
      llm mlx download-model mlx-community/gemma-3-27b-it-qat-4bit
    
      llm -m mlx-community/gemma-3-27b-it-qat-4bit \
        -f https://raw.githubusercontent.com/simonw/llm-hacker-news/refs/heads/main/llm_hacker_news.py \
        -f https://raw.githubusercontent.com/simonw/tools/refs/heads/main/github-issue-to-markdown.html \
        -s 'Write a new fragments plugin in Python that registers
        issue:org/repo/123 which fetches that issue
            number from the specified github repo and uses the same
            markdown logic as the HTML page to turn that into a
            fragment'
    
    It gave a solid response! https://gist.github.com/simonw/feccff6ce3254556b848c27333f52... - more notes here: https://simonwillison.net/2025/Apr/20/llm-fragments-github/
    • rs186 6 hours ago
      Can you quote tps?

      More and more I start to realize that cost saving is a small problem for local LLMs. If it is too slow, it becomes unusable, so much that you might as well use public LLM endpoints. Unless you really care about getting things done locally without sending information to another server.

      With OpenAI API/ChatGPT, I get response much faster than I can read, and for simple question, it means I just need a glimpse of the response, copy & paste and get things done. Whereas on local LLM, I watch it painstakingly prints preambles that I don't care about, and get what I actually need after 20 seconds (on a fast GPU).

      And I am not yet talking about context window etc.

      I have been researching about how people integrate local LLMs in their workflows. My finding is that most people play with it for a short time and that's about it, and most people are much better off spending money on OpenAI credits (which can last a very long time with typical usage) than getting a beefed up Mac Studio or building a machine with 4090.

      • simonw 6 hours ago
        My tooling doesn't measure TPS yet. It feels snappy to me on MLX.

        I agree that hosted models are usually a better option for most people - much faster, higher quality, handle longer inputs, really cheap.

        I enjoy local models for research and for the occasional offline scenario.

        I'm also interested in their applications for journalism, specifically for dealing with extremely sensitive data like leaked information from confidential sources.

        • freeamz 5 hours ago
          >I'm also interested in their applications for journalism, specifically for dealing with extremely sensitive data like leaked information from confidential sources.

          Think it is NOT just you. Most company with decent management also would not want their data going to anything outside the physical server they have in control of. But yeah for most people just use an app and hosted server. But this is HN,there are ppl here hosting their own email servers, so shouldn't be too hard to run llm locally.

          • simonw 5 hours ago
            "Most company with decent management also would not want their data going to anything outside the physical server they have in control of."

            I don't think that's been true for over a decade: AWS wouldn't be trillion dollar business if most companies still wanted to stay on-premise.

            • terhechte 4 hours ago
              Or GitHub. I’m always amused when people don’t want to send fractions of their code to a LLM but happily host it on GitHub. All big llm providers offer no-training-on-your-data business plans.
              • tarruda 4 hours ago
                > I’m always amused when people don’t want to send fractions of their code to a LLM but happily host it on GitHub

                What amuses me even more is people thinking their code is too unique and precious, and that GitHub/Microsoft wants to steal it.

                • vikarti 58 minutes ago
                  Regulations sometimes matter. Stupid "security" rules sometimes matter too.
                • AlexCoventry 4 hours ago
                  Concern about platform risk in regard to Microsoft is historically justified.
                • Terretta 4 hours ago
                  Unlikely they think Microsoft or GitHub wants to steal it.

                  With LLMs, they're thinking of examples that regurgitated proprietary code, and contrary to everyday general observation, valuable proprietary code does exist.

                  But with GitHub, the thinking is generally the opposite: the worry is that the code is terrible, and seeing it would be like giant blinkenlights* indicating the way in.

                  * https://en.wikipedia.org/wiki/Blinkenlights

            • __float 4 hours ago
              While none of that is false, I think there's a big difference from shipping your data to an external LLM API and using AWS.

              Using AWS is basically a "physical server they have control of".

              • simonw 3 hours ago
                That's why AWS Bedrock and Google Vertex AI and Azure AI model inference exist - they're all hosted LLM services that offer the same compliance guarantees that you get from regular AWS-style hosting agreements.
              • IanCal 2 hours ago
                As in aws is a much bigger security concern?
      • k__ 25 minutes ago
        The local LLM is your project manager, the big remote ones are the engineers and designers :D
      • overfeed 5 hours ago
        > Whereas on local LLM, I watch it painstakingly prints preambles that I don't care about, and get what I actually need after 20 seconds.

        You may need to "right-size" the models you use to match your hardware, model, and TPS expectations, which may involve using a smaller version of the model with faster TPS, upgrading your jardware, or paying for hosted models.

        Alternatively, if you can use agentic workflows or tools like Aider, you don't have to watch the model work slowly with large modles locally. Instead you queue work for it, go to sleep, or eat, or do other work, and then much later look over the Pull Requests whenever it completes them.

        • rs186 15 minutes ago
          I have a 4070 super for gaming, and used it to play with LLM a few times. It is by no means a bad card, but I realize that unless I want to get 4090 or new Macs that I don't have any other use for, I can only use it to run smaller models. However, most smaller models aren't satisfactory and are still slower than hosted LLMs. I haven't found a model that I am happy with for my hardware.

          Regarding agentic workflows -- sounds nice but I am too scared to try it out, based on my experience with standard LLMs like GPT or Claude for writing code. Small snippets or filling in missing unit tests, fine, anything more complicated? Has been a disaster for me.

      • DJHenk 2 hours ago
        > More and more I start to realize that cost saving is a small problem for local LLMs. If it is too slow, it becomes unusable, so much that you might as well use public LLM endpoints. Unless you really care about getting things done locally without sending information to another server.

        There is another aspect to consider, aside from privacy.

        These models are trained by downloading every scrap of information from the internet, including the works of many, many authors who have never consented to that. And they for sure are not going to get a share of the profits, if there is every going to be any. If you use a cloud provider, you are basically saying that is all fine. You are happy to pay them, and make yourself dependent on their service, based on work that wasn't theirs to use.

        However, if you use a local model, the authors still did not give consent, but one could argue that the company that made the model is at least giving back to the community. They don't get any money out of it, and you are not becoming dependent on their hyper capitalist service. No rent-seeking. The benefits of the work are free to use for everyone. This makes using AI a little more acceptable from a moral standpoint.

      • ein0p 2 hours ago
        Sometimes TPS doesn't matter. I've generated textual descriptions for 100K or so images in my photo archive, some of which I have absolutely no interest in uploading to someone else's computer. This works pretty well with Gemma. I use local LLMs all the time for things where privacy is even remotely important. I estimate this constitutes easily a quarter of my LLM usage.
        • lodovic 1 hour ago
          This is a really cool idea. Do you pretrain the model so it can tag people? I have so many photo's that it seems impossible to ever categorize them,using a workflow like yours might help a lot
          • ein0p 1 hour ago
            No, tagging of people is already handled by another model. Gemma just describes what's in the image, and produces a comma separated list of keywords. No additional training is required besides a few tweaks to the prompt so that it outputs just the description, without any "fluff". E.g. it normally prepends such outputs with "Here's a description of the image:" unless you really insist that it should output only the description. I suppose I could use constrained decoding into JSON or something to achieve the same, but I didn't mess with that.

            On some images where Gemma3 struggles Mistral Small produces better descriptions, BTW. But it seems harder to make it follow my instructions exactly.

            I'm looking forward to the day when I can also do this with videos, a lot of which I also have no interest in uploading to someone else's computer.

            • fer 18 minutes ago
              How do you use the keywords after? I have Immich running which does some analysis, but the querying is a bit of a hit and miss.
              • ein0p 4 minutes ago
                Search is indeed hit and miss. Immich, for instance, currently does absolutely nothing with the EXIF "description" field, so I store textual descriptions on the side as well. I have found Immich's search by image embeddings to be pretty weak at recall, and even weaker at ranking. IIRC Lightroom Classic (which I also use, but haven't found a way to automate this for without writing an extension) does search that field, but ranking is a bit of a dumpster fire, so your best bet is searching uncommon terms or constraining search by metadata (e.g. not just "black kitten" but "black kitten AND 2025"). I expect this to improve significantly over time - it's a fairly obvious thing to add given the available tech.
      • otabdeveloper4 4 hours ago
        The only actually useful application of LLM's is processing large amounts of data for classification and/or summarizing purposes.

        That's not the stuff you want to send to a public API, this is something you want as a 24/7 locally running batch job.

        ("AI assistant" is an evolutionary dead end, and Star Trek be damned.)

    • prvc 32 minutes ago
      > ~15GB (MLX) leaving plenty of memory for running other apps.

      Is that small enough to run well (without thrashing) on a system with only 16GiB RAM?

    • paprots 1 hour ago
      The original gemma3:27b also took only 22GB using Ollama on my 64GB MacBook. I'm quite confused that the QAT took the same. Do you know why? Which model is better? `gemma3:27b`, or `gemma3:27b-qat`?
      • kgwgk 9 minutes ago
        Look up 27b in https://ollama.com/library/gemma3/tags

        You'll find the id a418f5838eaf which also corresponds to 27b-it-q4_K_M

      • nolist_policy 1 hour ago
        I suspect your "original gemma3:27b" was a quantized model since the non-quantized (16bit) version needs around 54gb.
    • tomrod 5 hours ago
      Simon, what is your local GPU setup? (No doubt you've covered this, but I'm not sure where to dig up).
      • simonw 5 hours ago
        MacBook Pro M2 with 64GB of RAM. That's why I tend to be limited to Ollama and MLX - stuff that requires NVIDIA doesn't work for me locally.
        • Elucalidavah 3 hours ago
          > MacBook Pro M2 with 64GB of RAM

          Are there non-mac options with similar capabilities?

          • simonw 3 hours ago
            Yes, but I don't really know anything about those. https://www.reddit.com/r/LocalLLaMA/ is full of people running models on PCs with NVIDIA cards.

            The unique benefit of an Apple Silicon Mac at the moment is that the 64GB of RAM is available to both the GPU and the CPU at once. With other hardware you usually need dedicated separate VRAM for the GPU.

          • _neil 1 hour ago
            It’s not out yet, but the upcoming Framework desktop [0] is supposed to have a similar unified memory setup.

            [0] https://frame.work/desktop

    • nico 5 hours ago
      Been super impressed with local models on mac. Love that the gemma models have 128k token context input size. However, outputs are usually pretty short

      Any tips on generating long output? Like multiple pages of a document, a story, a play or even a book?

      • simonw 5 hours ago
        The tool you are using may set a default max output size without you realizing. Ollama has a num_ctx that defaults to 2048 for example: https://github.com/ollama/ollama/blob/main/docs/faq.md#how-c...
        • nico 5 hours ago
          Been playing with that, but doesn’t seem to have much effect. It works very well to limit output to smaller bits, like setting it to 100-200. But above 2-4k the output seems to never get longer than about 1 page

          Might try using the models with mlx instead of ollama to see if that makes a difference

          Any tips on prompting to get longer outputs?

          Also, does the model context size determine max output size? Are the two related or are they independent characteristics of the model?

          • simonw 5 hours ago
            Interestingly the Gemma 3 docs say: https://ai.google.dev/gemma/docs/core/model_card_3#:~:text=T...

            > Total output context up to 128K tokens for the 4B, 12B, and 27B sizes, and 32K tokens for the 1B size per request, subtracting the request input tokens

            I don't know how to get it to output anything that length though.

            • nico 4 hours ago
              Thank you for the insights and useful links

              Will keep experimenting, will also try mistral3.1

              edit: just tried mistral3.1 and the quality of the output is very good, at least compared to the other models I tried (llama2:7b-chat, llama2:latest, gemma3:12b, qwq and deepseek-r1:14b)

              Doing some research, because of their training sets, it seems like most models are not trained on producing long outputs so even if they technically could, they won’t. Might require developing my own training dataset and then doing some fine tuning. Apparently the models and ollama have some safeguards against rambling and repetition

              • Gracana 1 hour ago
                You can probably find some long-form tuned models on HF. I've had decent results with QwQ-32B (which I can run on my desktop) and Mistral Large (which I have to run on my server). Generating and refining an outline before writing the whole piece can help, and you can also split the piece up into multiple outputs (working a paragraph or two at a time, for instance). So far I've found it to be a tough process, with mixed results.
      • Casteil 5 hours ago
        This is basically the opposite of what I've experienced - at least compared to another recent entry like IBM's Granite 3.3.

        By comparison, Gemma3's output (both 12b and 27b) seems to typically be more long/verbose, but not problematically so.

        • nico 2 hours ago
          I agree with you. The outputs are usually good, it’s just that for the use case I have now (writing several pages of long dialogs), the output is not as long as I’d want it, and definitely not as long as it’s supposedly capable of doing
    • littlestymaar 3 hours ago
      > and it only uses ~22Gb (via Ollama) or ~15GB (MLX)

      Why is the memory use different? Are you using different context size in both set-ups?

  • diggan 7 hours ago
    First graph is a comparison of the "Elo Score" while using "native" BF16 precision in various models, second graph is comparing VRAM usage between native BF16 precision and their QAT models, but since this method is about doing quantization while also maintaining quality, isn't the obvious graph of comparing the quality between BF16 and QAT missing? The text doesn't seem to talk about it either, yet it's basically the topic of the blog post.
    • croemer 6 hours ago
      Indeed, the one thing I was looking for was Elo/performance of the quantized models, not how good the base model is. Showing how much memory is saved by quantization in a figure is a bit of an insult to the intelligence of the reader.
    • nithril 6 hours ago
      In addition the graph "Massive VRAM Savings" graph states what looks like a tautology, reducing from 16 bits to 4 bits leads unsurprisingly to a x4 reduction in memory usage
    • claiir 2 hours ago
      Yea they mention a “perplexity drop” relative to naive quantization, but that’s meaningless to me. > We reduce the perplexity drop by 54% (using llama.cpp perplexity evaluation) when quantizing down to Q4_0.

      Wish they showed benchmarks / added quantized versions to the arena! :>

  • Samin100 38 minutes ago
    I have a few private “vibe check” questions and the 4 bit QAT 27B model got them all correctly. I’m kind of shocked at the information density locked in just 13 GB of weights. If anyone at Deepmind is reading this — Gemma 3 27B is the single most impressive open source model I have ever used. Well done!
  • trebligdivad 6 hours ago
    It seems pretty impressive - I'm running it on my CPU (16 core AMD 3950x) and it's very very impressive at translation, and the image description is very impressive as well. I'm getting about 2.3token/s on it (compared to under 1/s on the Calme-3.2 I was previously using). It does tend to be a bit chatty unless you tell it not to be; pretty much everything it'll give you a 'breakdown' unless you tell it not to - so for traslation my prompt is 'Translate the input to English, only output the translation' to stop it giving a breakdown of the input language.
    • simonw 6 hours ago
      What are you using to run it? I haven't got image input working yet myself.
      • trebligdivad 6 hours ago
        I'm using llama.cpp - built last night from head; to do image stuff you have to run a separate client they provide, with something like:

        ./build/bin/llama-gemma3-cli -m /discs/fast/ai/gemma-3-27b-it-q4_0.gguf --mmproj /discs/fast/ai/mmproj-model-f16-27B.gguf -p "Describe this image." --image ~/Downloads/surprise.png

        Note the 2nd gguf in there - I'm not sure, but I think that's for encoding the image.

      • terhechte 4 hours ago
        Image input has been working with LM Studio for quite some time
  • mythz 7 hours ago
    The speed gains are real, after downloading latest QAT gemma3:27b eval perf is now 1.47x faster on ollama, up from 13.72 to 20.11 tok/s (on A4000's).
  • porphyra 3 hours ago
    It is funny that Microsoft had been peddling "AI PCs" and Apple had been peddling "made for Apple Intelligence" for a while now, when in fact usable models for consumer GPUs are only barely starting to be a thing on extremely high end GPUs like the 3090.
    • ivape 1 hour ago
      This is why the "AI hardware cycle is hype" crowd is so wrong. We're not even close, we're basically at ColecoVision/Atari stage of hardware here. It's going be quite a thing when everyone gets a SNES/Genesis.
  • mark_l_watson 1 hour ago
    Indeed!! I have swapped out qwen2.5 for gemma3:27b-it-qat using Ollama for routine work on my 32G memory Mac.

    gemma3:27b-it-qat with open-codex, running locally, is just amazingly useful, not only for Python dev, but for Haskell and Common Lisp also.

    I still like Gemini 2.5 Pro and o3 for brainstorming or working on difficult problems, but for routine work it (simply) makes me feel good to have everything open source/weights running on my own system.

    Wen I bought my 32G Mac a year ago, I didn't expect to be so happy as running gemma3:27b-it-qat with open-codex locally.

  • behnamoh 7 hours ago
    This is what local LLMs need—being treated like first-class citizens by the companies that make them.

    That said, the first graph is misleading about the number of H100s required to run DeepSeek r1 at FP16. The model is FP8.

    • mmoskal 1 hour ago
      Also ~noone runs h100 at home, ie at batch size 1. What matters is throughput. With 37b active parameters and a massive deployment throughout (per gpu) should be similar to Gemma.
    • freeamz 5 hours ago
      so what is the real comparison against DeepSeek r1 ? Would be good to know which is actually more cost efficient and open (reproducible build) to run locally.
      • behnamoh 5 hours ago
        half the amount of those dots is what it takes. but also, why compare a 27B model with a +600B? that doesn't make sense.
  • mekpro 5 hours ago
    Gemma 3 is way way better than Llama 4. I think Meta will start to lose its position in LLM mindshare. Another weakness of Llama 4 is its model size that is too large (even though it can run fast with MoE), which greatly limits the applicable users to a small percentage of enthusiasts who have enough GPU VRAM. Meanwhile, Gemma 3 is widely usable across all hardware sizes.
  • emrah 8 hours ago
    • jinay 7 hours ago
      Make sure you're using the "-it-qat" suffixed models like "gemma3:27b-it-qat"
    • Der_Einzige 7 hours ago
      How many times do I have to say this? Ollama, llamacpp, and many other projects are slower than vLLM/sglang. vLLM is a much superior inference engine and is fully supported by the only LLM frontends that matter (sillytavern).

      The community getting obsessed with Ollama has done huge damage to the field, as it's ineffecient compared to vLLM. Many people can get far more tok/s than they think they could if only they knew the right tools.

      • Zambyte 6 hours ago
        The significant convenience benefits outweigh the higher TPS that vLLM offers in the context of my single machine homelab GPU server. If I was hosting it for something more critical than just myself and a few friends chatting with it, sure. Being able to just paste a model name into Open WebUI and run it is important to me though.

        It is important to know about both to decide between the two for your use case though.

        • Der_Einzige 3 hours ago
          Running any HF model on vllm is as simple as pasting a model name into one command in your terminal.
          • Zambyte 1 hour ago
            What command is it? Because that was not at all my experience.
      • ach9l 6 hours ago
        instead of ranting, maybe explain how to make a qat q4 work with images in vllm, afaik it is not yet possible
      • simonw 6 hours ago
        Last I looked vLLM didn't work on a Mac.
        • mitjam 4 hours ago
          Afaik vllm is for concurrent serving with batched inference for higher throughput, not single-user inference. I doubt inference throughput is higher with single prompts at a time than Ollama. Update: this is a good Intro to continuous batching in llm inference: https://www.anyscale.com/blog/continuous-batching-llm-infere...
          • Der_Einzige 3 hours ago
            It is much faster on single prompts than ollama. 3X is not unheard of
      • oezi 6 hours ago
        Why is sillytavern the only LLM frontend which matters?
        • GordonS 6 hours ago
          I tried sillytavern a few weeks ago... wow, that is an "interesting" UI! I blundered around for a while, couldn't figure out how to do anything useful... and then installed LM Studio instead.
          • imtringued 5 hours ago
            I personally thought the lorebook feature was quite neat and then quickly gave up on it because I couldn't get it to trigger, ever.

            Whatever those keyword things are, they certainly don't seem to be doing any form of RAG.

        • Der_Einzige 3 hours ago
          It supports more sampler and other settings than anyone else.
      • janderson215 7 hours ago
        I did not know this, so thank you. I read a blogpost a while back that encouraged using Ollama and never mention vLLM. Do you recommend reading any particular resource?
      • oezi 6 hours ago
        Somebody in this thread mentioned 20.x tok/s on ollama. What are you seeing in vLLM?
        • Zambyte 6 hours ago
          FWIW I'm getting 29 TPS on Ollama on my 7900 XTX with the 27b qat. You can't really compare inference engine to inference engine without keeping the hardware and model fixed.

          Unfortunately Ollama and vLLM are therefore incomparable at the moment, because vLLM does not support these models yet.

          https://github.com/vllm-project/vllm/issues/16856

      • m00dy 7 hours ago
        Ollama is definitely not for production loads but vLLm is.
  • cheriot 19 minutes ago
    Is there already a Helium for GPUs?
  • miki123211 4 hours ago
    What would be the best way to deploy this if you're maximizing for GPU utilization in a multi-user (API) scenario? Structured output support would be a big plus.

    We're working with a GPU-poor organization with very strict data residency requirements, and these models might be exactly what we need.

    I would normally say VLLM, but the blog post notably does not mention VLLM support.

  • piyh 1 hour ago
    Meta Maverick is crying in the shower getting so handily beat by a model with 15x fewer params
  • 999900000999 4 hours ago
    Assuming this can match Claude's latest, and full time usage ( as in you have a system that's constantly running code without any user input,) you'd probably save 600 to 700 a month. A 4090 is only 2K and you'll see an ROI within 90 days.

    I can imagine this will serve to drive prices for hosted llms lower.

    At this level any company that produces even a nominal amount of code should be running LMS on prem( AWS if your on the cloud).

  • umajho 6 hours ago
    I am currently using the Q4_K_M quantized version of gemma-3-27b-it locally. I previously assumed that a 27B model with image input support wouldn't be very high quality, but after actually using it, the generated responses feel better than those from my previously used DeepSeek-R1-Distill-Qwen-32B (Q4_K_M), and its recognition of images is also stronger than I expected. (I thought the model could only roughly understand the concepts in the image, but I didn't expect it to be able to recognize text within the image.)

    Since this article publishes the optimized Q4 quantized version, it would be great if it included more comparisons between the new version and my currently used unoptimized Q4 version (such as benchmark scores).

    (I deliberately wrote this reply in Chinese and had gemma-3-27b-it Q4_K_M translate it into English.)

  • holografix 7 hours ago
    Could 16gb vram be enough for the 27b QAT version?
    • jffry 7 hours ago
      With `ollama run gemma3:27b-it-qat "What is blue"`, GPU memory usage is just a hair over 20GB, so no, probably not without a nerfed context window
      • woadwarrior01 6 hours ago
        Indeed, the default context length in ollama is a mere 2048 tokens.
    • hskalin 6 hours ago
      With ollama you could offload a few layers to cpu if they don't fit in the VRAM. This will cost some performance ofcourse but it's much better than the alternative (everything on cpu)
      • senko 4 hours ago
        I'm doing that with a 12GB card, ollama supports it out of the box.

        For some reason, it only uses around 7GB of VRAM, probably due to how the layers are scheduled, maybe I could tweak something there, but didn't bother just for testing.

        Obviously, perf depends on CPU, GPU and RAM, but on my machine (3060 + i5-13500) it's around 2 t/s.

    • halflings 7 hours ago
      That's what the chart says yes. 14.1GB VRAM usage for the 27B model.
      • erichocean 7 hours ago
        That's the VRAM required just to load the model weights.

        To actually use a model, you need a context window. Realistically, you'll want a 20GB GPU or larger, depending on how many tokens you need.

        • oezi 6 hours ago
          I didn't realize that the context would require such so much memory. Is this KV caches? It would seem like a big advantage if this memory requirement could be reduced.
    • parched99 5 hours ago
      I am only able to get the Gemma-3-27b-it-qat-Q4_0.gguf (15.6GB) to run with a 100 token context size on a 5070 ti (16GB) using llamacpp.

      Prompt Tokens: 10

      Time: 229.089 ms

      Speed: 43.7 t/s

      Generation Tokens: 41

      Time: 959.412 ms

      Speed: 42.7 t/s

      • tbocek 1 hour ago
        This is probably due to this: https://github.com/ggml-org/llama.cpp/issues/12637. This GitHub issue is about interleaved sliding window attention (iSWA) not available in llama.cpp for Gemma 3. This could reduce the memory requirements a lot. They mentioned for a certain scenario, going from 62GB to 10GB.
        • parched99 23 minutes ago
          Resolving that issue, would help reduce (not eliminate) the size of the context. The model will still only just barely fit in 16 GB, which is what the parent comment asked.

          Best to have two or more low-end, 16GB GPUs for a total of 32GB VRAM to run most of the better local models.

      • floridianfisher 1 hour ago
        Try one of the smaller versions. 27b is too big for your gpu
        • parched99 44 minutes ago
          I'm aware. I was addressing the question being asked.
  • noodletheworld 7 hours ago
    ?

    Am I missing something?

    These have been out for a while; if you follow the HF link you can see, for example, the 27b quant has been downloaded from HF 64,000 times over the last 10 days.

    Is there something more to this, or is just a follow up blog post?

    (is it just that ollama finally has partial (no images right?) support? Or something else?)

    • deepsquirrelnet 7 hours ago
      QAT “quantization aware training” means they had it quantized to 4 bits during training rather than after training in full or half precision. It’s supposedly a higher quality, but unfortunately they don’t show any comparisons between QAT and post-training quantization.
      • noodletheworld 7 hours ago
        I understand that, but the qat models (1) are not new uploads.

        How is this more significant now than when they were uploaded 2 weeks ago?

        Are we expecting new models? I don’t understand the timing. This post feels like it’s two weeks late.

        [1] - https://huggingface.co/collections/google/gemma-3-qat-67ee61...

        • simonw 6 hours ago
          The official announcement of the QAT models happened on Friday 18th, two days ago. It looks like they uploaded them to HF in advance of that announcement: https://developers.googleblog.com/en/gemma-3-quantized-aware...

          The partnership with Ollama and MLX and LM Studio and llama.cpp was revealed in that announcement, which made the models a lot easier for people to use.

        • llmguy 6 hours ago
          8 days is closer to 1 week then 2. And it’s a blog post, nobody owes you realtime updates.
          • noodletheworld 6 hours ago
            https://huggingface.co/google/gemma-3-27b-it-qat-q4_0-gguf/t...

            > 17 days ago

            Anywaaay...

            I'm literally asking, quite honestly, if this is just an 'after the fact' update literally weeks later, that they uploaded a bunch of models, or if there is something more significant about this I'm missing.

            • osanseviero 5 hours ago
              Hi! Omar from the Gemma team here.

              Last time we only released the quantized GGUFs. Only llama.cpp users could use it (+ Ollama, but without vision).

              Now, we released the unquantized checkpoints, so anyone can quantize themselves and use in their favorite tools, including Ollama with vision, MLX, LM Studio, etc. MLX folks also found that the model worked decently with 3 bits compared to naive 3-bit, so by releasing the unquantized checkpoints we allow further experimentation and research.

              TL;DR. One was a release in a specific format/tool, we followed-up with a full release of artifacts that enable the community to do much more.

              • oezi 5 hours ago
                Hey Omar, is there any chance that Gemma 3 might get a speech (ASR/AST/TTS) release?
            • timcobb 6 hours ago
              Probably the former... I see your confusion but it's really only a couple weeks at most. The news cycle is strong in you, grasshopper :)
    • xnx 7 hours ago
      The linked blog post was 2 days ago
  • justanotheratom 6 hours ago
    Anyone packaged one of these in an iPhone App? I am sure it is doable, but I am curious what tokens/sec is possible these days. I would love to ship "private" AI Apps if we can get reasonable tokens/sec.
    • zamadatix 5 hours ago
      There are many such apps, e.g. Mollama, Enclave AI or PrivateLLM or dozens of others, but you could tell me it runs at 1,000,000 tokens/second on an iPhone and I wouldn't care because the largest model version you're going to be able to load is Gemma 3 4B q4 (12 B won't fit in 8 GB with the OS + you still need context) and it's just not worth the time to use.

      That said, if you really care, it generates faster than reading speed (on an A18 based model at least).

      • woodson 5 hours ago
        Some of these small models still have their uses, e.g. for summarization. Don’t expect them to fully replace ChatGPT.
        • zamadatix 4 hours ago
          The use case is more "I'm willing to have really bad answers that have extremely high rates of making things up" than based on the application. The same goes for summarization, it's not like it does it well like a large model would.
    • Alifatisk 6 hours ago
      If you ever ship a private AI app, don't forget to implement the export functionality, please!
    • nolist_policy 4 hours ago
      FWIW, I can run Gemma-3-12b-it-qat on my Galaxy Fold 4 with 12Gb ram at around 1.5 tokens / s. I use plain llama.cpp with Termux.
      • Casteil 3 hours ago
        Does this turn your phone into a personal space heater too?
    • nico 5 hours ago
      What kind of functionality do you need from the model?

      For basic conversation and RAG, you can use tinyllama or qwen-2.5-0.5b, both of which run on a raspberry pi at around 5-20 tokens per second

  • CyberShadow 5 hours ago
    How does it compare to CodeGemma for programming tasks?
  • Alifatisk 6 hours ago
    Except this being lighter than the other models, is there anything else the Gemma model is specifically good at or better than the other models at doing?
    • Zambyte 5 hours ago
      I have found Gemma models are able to produce useful information about more niche subjects that other models like Mistral Small cannot, at the expense of never really saying "I don't know", where other models will, and will instead produce false information.

      For example, if I ask mistral small who I am by name, it will say there is no known notable figure by that name before the knowledge cutoff. Gemma 3 will say I am a well known <random profession> and make up facts. On the other hand, I have asked both about local organization in my area that I am involved with, and Gemma 3 could produce useful and factual information, where Mistral Small said it did not know.

    • nico 5 hours ago
      They are multimodal. Havent tried the QAT one yet. But the gemma3s released a few weeks ago are pretty good at processing images and telling you details about what’s in them
    • itake 6 hours ago
      Google claims to have better multi language support, due tokenizer improvements.
  • jarbus 7 hours ago
    Very excited to see these kinds of techniques, I think getting a 30B level reasoning model usable on consumer hardware is going to be a game changer, especially if it uses less power.
    • apples_oranges 7 hours ago
      Deepseek does reasoning on my home Linux pc but not sure how power hungry it is
      • gcr 7 hours ago
        what variant? I’d considered DeepSeek far too large for any consumer GPUs
        • scosman 7 hours ago
          Some people run Deepseek on CPU. 37B active params - it isn't fast but it's passible.
          • danielbln 6 hours ago
            Actual deepseek or some qwen/llama reasoning fine-tune?
            • scosman 4 hours ago
              Actual Deepseek. 500gb of memory and a threadripper works. Not a standard PC spec, but a common ish home brew setup for single user Deepseek.
  • wtcactus 7 hours ago
    They keep mentioning the RTX 3090 (with 24 GB VRAM), but the model is only 14.1 GB.

    Shouldn’t it fit a 5060 Ti 16GB, for instance?

    • oktoberpaard 6 hours ago
      With a 128K context length and 8 bit KV cache, the 27b model occupies 22 GiB on my system. With a smaller context length you should be able to fit it on a 16 GiB GPU.
    • jsnell 7 hours ago
      Memory is needed for more than just the parameters, e.g. the KV cache.
  • btbuildem 6 hours ago
    Is 27B the largest QAT Gemma 3? Given these size reductions, it would be amazing to have the 70B!
    • arnaudsm 6 hours ago
      The original Gemma 3 does not have a 70B version.
  • XCSme 5 hours ago
    So how does 27b-it-qat (18GB) compare to 27b-it-q4_K_M (17GB)?
  • briandear 3 hours ago
    The normal Gemma models seem to work fine on Apple silicon with Metal. Am I missing something?
    • simonw 3 hours ago
      These new special editions of those models claim to work better with less memory.
  • perching_aix 5 hours ago
    This is my first time trying to locally host a model - gave both the 12B and 27B QAT models a shot.

    I was both impressed and disappointed. Setup was piss easy, and the models are great conversationalists. I have a 12 gig card available and the 12B model ran very nice and swift.

    However, they're seemingly terrible at actually assisting with stuff. Tried something very basic: asked for a powershell one liner to get the native blocksize of my disks. Ended up hallucinating fields, then telling me to go off into the deep end, first elevating to admin, then using WMI, then bringing up IOCTL. Pretty unfortunate. Not sure I'll be able to put it to actual meaningful use as a result.

    • terhechte 4 hours ago
      Local models, due to their size more than big cloud models, favor popular languages rather than more niche ones. They work fantastic for JavaScript, Python, Bash but much worse at less popular things like Clojure, Nim or Haskell. Powershell is probably on the less popular side compared to Js or Bash.

      If this is your main use case you can always try to fine tune a model. I maintain a small llm bench of different programming languages and the performance difference between say Python and Rust on some smaller models is up to 70%

      • perching_aix 4 hours ago
        How accessible and viable is model fine-tuning? I'm not in the loop at all unfortunately.
    • parched99 4 hours ago
      I think Powershell is a bad test. I've noticed all local models have trouble providing accurate responses to Powershell-related prompts. Strangely, even Microsoft's model, Phi 4, is bad at answering these questions without careful prompting. Though, MS can't even provide accurate PS docs.

      My best guess is that there's not enough discussion/development related to Powershell in training data.

      • fragmede 30 minutes ago
        Which, like, you'd think Microsoft has an entire team there who's purpose would be to generate good PowerShell for it to train on.
  • api 5 hours ago
    When I see 32B or 70B models performing similarly to 200+B models, I don’t know what to make of this. Either the latter contains more breadth of information but we have managed to distill latent capabilities to be similar, the larger models are just less efficient, or the tests are not very good.
    • simonw 4 hours ago
      It makes intuitive sense to me that this would be possible, because LLMs are still mostly opaque black boxes. I expect you could drop a whole hunch of the weights without having a huge impact on quality - maybe you end up mostly ditching the parts that are derived from shitposts on Reddit but keep the bits from Arxiv for example.

      (That's a massive simplification of how any of this works, but it's how I think about it at a high level.)

    • retinaros 4 hours ago
      its just bs benchmarks. they are all cheating at this point feeding the data in the training set. doesnt mean the llm arent becoming better but when they all lie...
  • rob_c 6 hours ago
    Given how long between this being released and this community picking up on it... Lol
    • GaunterODimm 2 hours ago
      2days :/...
      • rob_c 2 hours ago
        Given I know people running gemma3 on local devices for over almost a month now this is either a very slow news day or evidence of finger missing the pulse... https://blog.google/technology/developers/gemma-3/
        • simonw 2 hours ago
          This is new. These are new QAT (Quantization-Aware Training) models released by the Gemma team.
          • rob_c 2 hours ago
            There's nothing more than an iteration on the topic, gemma3 was smashing local results a month ago and made no waves as it dropped...
            • simonw 2 hours ago
              Quoting the linked story:

              > Last month, we launched Gemma 3, our latest generation of open models. Delivering state-of-the-art performance, Gemma 3 quickly established itself as a leading model capable of running on a single high-end GPU like the NVIDIA H100 using its native BFloat16 (BF16) precision.

              > To make Gemma 3 even more accessible, we are announcing new versions optimized with Quantization-Aware Training (QAT) that dramatically reduces memory requirements while maintaining high quality.

              The thing that's new, and that is clearly resonating with people, is the "To make Gemma 3 even more accessible..." bit.

              • rob_c 2 hours ago
                As I've said in my lectures on how to perform 1bit training of QAT systems to build classifiers...

                "An iteration on a theme".

                Once the network design is proven to work yes it's an impressive technical achievement, but as I've said given I've known people in multiple research institutes and companies using Gemma3 for a month mostly saying they're surprised it's not getting noticed... This is just enabling more users but the none QAT version will almost always perform better...

                • simonw 2 hours ago
                  Sounds like you're excited to see Gemma 3 get the recognition it deserves on Hacker News then.
                  • rob_c 2 hours ago
                    No just pointing out the flooding obvious as usual and collecting down votes for it
                    • fragmede 17 minutes ago
                      Speaking for myself, my downvotes are not because of the content of your arguments, but because your tone is consistently condescending and dismissive. Comments like “just pointing out the flooding obvious” come off as smug and combative rather than constructive.

                      HN works best when people engage in good faith, stay curious, and try to move the conversation forward. That kind of tone — even when technically accurate — discourages others from participating and derails meaningful discussion.

                      If you’re getting downvotes regularly, maybe it's worth considering how your comments are landing with others, not just whether they’re “right.”