frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Google releases Gemma 4 open models

https://deepmind.google/models/gemma/gemma-4/
456•jeffmcjunkin•1h ago•118 comments

Qwen3.6-Plus: Towards real world agents

https://qwen.ai/blog?id=qwen3.6
243•pretext•3h ago•86 comments

LinkedIn is illegally searching your computer

https://browsergate.eu/
1158•digitalWestie•4h ago•535 comments

Lemonade by AMD: a fast and open source local LLM server using GPU and NPU

https://lemonade-server.ai
298•AbuAssar•7h ago•76 comments

A forecast of the fair market value of SpaceX's businesses

https://futuresearch.ai/spacex-ipo-valuation/
72•ddp26•1h ago•71 comments

Even GPT-5.2 Can't Count to Five: Zero-Error Horizons in Trustworthy LLMs

https://arxiv.org/abs/2601.15714
26•daigoba66•2h ago•19 comments

Inside Nepal's Fake Rescue Racket

https://kathmandupost.com/money/2026/03/27/inside-nepal-s-fake-rescue-racket
190•lode•6h ago•76 comments

Delve allegedly forked an open-source tool and sold it as its own

https://techcrunch.com/2026/04/01/the-reputation-of-troubled-yc-startup-delve-has-gotten-even-worse/
166•nickvec•3h ago•81 comments

Significant Raise of Reports

https://lwn.net/Articles/1065620/
218•stratos123•8h ago•102 comments

Artemis II will use laser beams to live-stream 4K moon footage at 260 Mbps

https://www.tomshardware.com/networking/artemis-ii-will-use-laser-beams-to-live-stream-4k-moon-fo...
177•speckx•3h ago•75 comments

IBM Announces Strategic Collaboration with Arm

https://newsroom.ibm.com/2026-04-02-ibm-announces-strategic-collaboration-with-arm-to-shape-the-f...
228•bonzini•9h ago•147 comments

Sweden goes back to basics, swapping screens for books in the classroom

https://undark.org/2026/04/01/sweden-schools-books/
571•novaRom•7h ago•297 comments

Renewables reached nearly 50% of global electricity capacity last year

https://www.theregister.com/2026/04/01/renewables_generated_nearly_half_global_power/
105•Growtika•2h ago•48 comments

OpenAI Acquires TBPN

https://openai.com/index/openai-acquires-tbpn/
23•surprisetalk•41m ago•13 comments

'Backrooms' and the Rise of the Institutional Gothic

https://thereader.mitpress.mit.edu/backrooms-and-the-rise-of-the-institutional-gothic/
95•anarbadalov•4h ago•45 comments

An Example of Statistical Investigation of the Text Eugene Onegin – Markov, 1913 [pdf]

https://alpha60.de/research/markov/DavidLink_AnExampleOfStatistical_MarkovTrans_2007.pdf
15•jxmorris12•3d ago•1 comments

Modern SQLite: Features You Didn't Know It Had

https://slicker.me/sqlite/features.htm
57•thunderbong•1h ago•5 comments

EmDash: A Fresh Take on CMS

https://maciekpalmowski.dev/blog/emdash-a-fresh-take-on-cms/
31•taubek•1h ago•9 comments

Bringing Clojure programming to Enterprise (2021)

https://blogit.michelin.io/clojure-programming/
161•smartmic•9h ago•95 comments

Gone (Almost) Phishin'

https://ma.tt/2026/03/gone-almost-phishin/
139•luu•2d ago•67 comments

Quadratic Micropass Type Inference

https://articles.luminalang.com/a/micropass-inference/
11•simvux•5d ago•4 comments

Email obfuscation: What works in 2026?

https://spencermortensen.com/articles/email-obfuscation/
295•jaden•14h ago•86 comments

Reinventing the pull request

https://lubeno.dev/blog/reinventing-the-pull-request
65•bkolobara•6d ago•51 comments

March heat in American west has left snowpack at record-low levels

https://www.theguardian.com/us-news/2026/apr/01/snowmelt-american-west
65•ijidak•2h ago•25 comments

Mercor says it was hit by cyberattack tied to compromise LiteLLM

https://techcrunch.com/2026/03/31/mercor-says-it-was-hit-by-cyberattack-tied-to-compromise-of-ope...
128•jackson-mcd•1d ago•36 comments

Emacs-libgterm: Terminal emulator for Emacs using libghostty-vt

https://github.com/rwc9u/emacs-libgterm
63•signa11•4d ago•16 comments

Telli (YC F24) is hiring engineers, designers, and more (on-site, Berlin)

http://hi.telli.com/join-us
1•sebselassie•10h ago

Show HN: I built a DNS resolver from scratch in Rust – no DNS libraries

https://github.com/razvandimescu/numa
65•rdme•7h ago•45 comments

Enabling Codex to Analyze Two Decades of Hacker News Data

https://modolap.com/publication/hn-analysis-1
59•ronfriedhaber•7h ago•25 comments

Steam on Linux Use Skyrocketed Above 5% in March

https://www.phoronix.com/news/Steam-On-Linux-Tops-5p
674•hkmaxpro•14h ago•306 comments
Open in hackernews

Google releases Gemma 4 open models

https://deepmind.google/models/gemma/gemma-4/
448•jeffmcjunkin•1h ago

Comments

danielhanchen•1h ago
Thinking / reasoning + multimodal + tool calling.

We made some quants at https://huggingface.co/collections/unsloth/gemma-4 for folks to run them - they work really well!

Guide for those interested: https://unsloth.ai/docs/models/gemma-4

Also note to use temperature = 1.0, top_p = 0.95, top_k = 64 and the EOS is "<turn|>". "<|channel>thought\n" is also used for the thinking trace!

l2dy•1h ago
FYI, screenshot for the "Search and download Gemma 4" step on your guide is for qwen3.5, and when I searched for gemma-4 in Unsloth Studio it only shows Gemma 3 models.
danielhanchen•1h ago
We're still updating it haha! Sorry! It's been quite complex to support new models without breaking old ones
Imustaskforhelp•1h ago
Daniel, I know you might hear this a lot but I really appreciate a lot of what you have been doing at Unsloth and the way you handle your communication, whether within hackernews/reddit.

I am not sure if someone might have asked this already to you, but I have a question (out of curiosity) as to which open source model you find best and also, which AI training team (Qwen/Gemini/Kimi/GLM) has cooperated the most with the Unsloth team and is friendly to work with from such perspective?

danielhanchen•1h ago
Thanks a lot for the support :)

Tbh Gemma-4 haha - it's sooooo good!!!

For teams - Google haha definitely hands down then Qwen, Meta haha through PyTorch and Llama and Mistral - tbh all labs are great!

Imustaskforhelp•1h ago
Now you have gotten me a bit excited for Gemma-4, Definitely gonna see if I can run the unsloth quants of this on my mac air & thanks for responding to my comment :-)
danielhanchen•1h ago
Thanks! Have a super good day!!
evilelectron•58m ago
Daniel, your work is changing the world. More power to you.

I setup a pipeline for inference with OCR, full text search, embedding and summarization of land records dating back 1800s. All powered by the GGUF's you generate and llama.cpp. People are so excited that they can now search the records in multiple languages that a 1 minute wait to process the document seems nothing. Thank you!

danielhanchen•57m ago
Oh appreciate it!

Oh nice! That sounds fantastic! I hope Gemma-4 will make it even better! The small ones 2B and 4B are shockingly good haha!

zaat•26m ago
Thank you for your work.

You have an answer on your page regarding "Should I pick 26B-A4B or 31B?", but can you please clarify if, assuming 24GB vRAM, I should pick a full precision smaller model or 4 bit larger model?

danielhanchen•7m ago
Thank you!

I presume 24B is somewhat faster since it's only 4B activated - 31B is quite a large dense model so more accurate!

jwr•1h ago
Really looking forward to testing and benchmarking this on my spam filtering benchmark. gemma-3-27b was a really strong model, surpassed later by gpt-oss:20b (which was also much faster). qwen models always had more variance.
jeffbee•1h ago
Does spam filtering really need a better model? My impression is that the whole game is based on having the best and freshest user-contributed labels.
a7om_com•1h ago
Gemma models are already in our AIPI inference pricing index. Open source models like Gemma run 70.7% cheaper than proprietary equivalents at the median across the 2,614 SKUs we track. With Gemma 4 hitting third-party platforms the pricing will be worth watching closely. Full data at a7om.com.
flakiness•1h ago
It's good they still have non-instruction-tuned models.
minimaxir•1h ago
The benchmark comparisons to Gemma 3 27B on Hugging Face are interesting: The Gemma 4 E4B variant (https://huggingface.co/google/gemma-4-E4B-it) beats the old 27B in every benchmark at a fraction of parameters.

The E2B/E4B models also support voice input, which is rare.

regularfry•53m ago
Thinking vs non-thinking. There'll be a token cost there. But still fairly remarkable!
DoctorOetker•25m ago
Is there a reason we can't use thinking completions to train non-thinking? i.e. gradient descent towards what thinking would have answered?
NitpickLawyer•1h ago
Best thing is that this is Apache 2.0 (edit: and they have base models available. Gemma3 was good for finetuning)

The sizes are E2B and E4B (following gemma3n arch, with focus on mobile) and 26BA4 MoE and 31B dense. The mobile ones have audio in (so I can see some local privacy focused translation apps) and the 31B seems to be strong in agentic stuff. 26BA4 stands somewhere in between, similar VRAM footprint, but much faster inference.

babelfish•1h ago
Wow, 30B parameters as capable as a 1T parameter model?
mhitza•14m ago
On the above compared benchmarks is closer to other larger open weights models, and on par with GPT-OSS 120B, for which I also have a frame of reference.
darshanmakwana•1h ago
This is awesome! I will try to use them locally with opencode and see if they are usable inreplacement of claude code for basic tasks
antirez•1h ago
Featuring the ELO score as the main benchmark in chart is very misleading. The big dense Gemma 4 model does not seem to reach Qwen 3.5 27B dense model in most benchmarks. This is obviously what matters. The small 2B / 4B models are interesting and may potentially be better ASR models than specialized ones (not just for performances but since they are going to be easily served via llama.cpp / MLX and front-ends). Also interesting for "fast" OCR, given they are vision models as well. But other than that, the release is a bit disappointing.
nabakin•1h ago
Public benchmarks can be trivially faked. Lmarena is a bit harder to fake and is human-evaluated.

I agree it's misleading for them to hyper-focus on one metric, but public benchmarks are far from the only thing that matters. I place more weight on Lmarena scores and private benchmarks.

moffkalast•26m ago
Lm arena is so easy to game that it's ceased to be a relevant metric over a year ago. People are not usable validators beyond "yeah that looks good to me", nobody checks if the facts are correct or not.
nabakin•7m ago
It's easy to game and human evaluation data has its trade-offs, but it's way easier to fake public benchmark results. I wish we had a source of high quality private benchmark results across a vast number of models like Lmarena. Having high quality human evaluation data would be a plus too.
jug•4m ago
I agree; LMArena died for me with the Llama 4 debacle. And not only the gamed scores, but seeing with shock and horror the answers people found good.
WarmWash•1h ago
I am unable to shake that the Chinese models all perform awfully on the private arc-agi 2 tests.
minimaxir•55m ago
I can't find what ELO score specifically the benchmark chart is referring to, it's just labeled "Elo Score". It's not Codeforces ELO as that Gemma 4 31B has 2150 for that which would be off the given chart.
nabakin•50m ago
It's referring to the Lmsys Leaderboard/Lmarena/Arena.ai[0]. It's very well-known in the LLM community for being one of the few sources of human evaluation data.

[0] https://arena.ai/leaderboard/chat

azinman2•55m ago
I find the benchmarks to be suggestive but not necessarily representative of reality. It's really best if you have your own use case and can benchmark the models yourself. I've found the results to be surprising and not what these public benchmarks would have you believe.
rvz•1h ago
Open weight models once again marching on and slowly being a viable alternative to the larger ones.

We are at least 1 year and at most 2 years until they surpass closed models for everyday tasks that can be done locally to save spending on tokens.

echelon•1h ago
> We are at least 1 year and at most 2 years until they surpass closed models for everyday tasks that can be done locally to save spending on tokens.

Until they pass what closed models today can do.

By that time, closed models will be 4 years ahead.

Google would not be giving this away if they believed local open models could win.

Google is doing this to slow down Anthropic, OpenAI, and the Chinese, knowing that in the fullness of time they can be the leader. They'll stop being so generous once the dust settles.

pixl97•1h ago
I mean, correct, but running open models locally will still massively drop your costs even if you still need to interface with large paid for models. Google will still make less money than if they were the only model that existed at the end of the day.
ma2kx•41m ago
I think it will be less of a local versus cloud situation, but rather one where both complement each other. The next step will undoubtedly be for local LLMs to be fast and intelligent enough to allow for vocal conversation. A low-latency model will then run locally, enabling smoother conversations, while batch jobs in the cloud handle the more complex tasks.

Google, at least, is likely interested in such a scenario, given their broad smartphone market. And if their local Gemma/Gemini-nano LLMs perform better with Gemini in the cloud, that would naturally be a significant advantage.

jimbokun•2m ago
But at that point, won’t there be very few tasks left where the average user can discern the difference in quality for most tasks?
james2doyle•1h ago
Hmm just tried the google/gemma-4-31B-it through HuggingFace (inference provider seems to be Novita) and function/tool calling was not enabled...
james2doyle•1h ago
Yeah you can see here that tool calling is disabled: https://huggingface.co/inference/models?model=google%2Fgemma...

At least, as of this post

linolevan•1h ago
Hosted on Parasail + Google (both for free, as of now) themselves, probably would give those a shot
originalvichy•1h ago
The wait is finally over. One or two iterations, and I’ll be happy to say that language models are more than fulfilling my most common needs when self-hosting. Thanks to the Gemma team!
adamtaylor_13•1h ago
What sort of tasks are you using self-hosting for? Just curious as I've been watching the scene but not experimenting with self-hosting.
irishcoffee•1h ago
I would personally be much more interested in using LLMs if I didn’t need to depend on an internet connection and spending money on tokens.
vunderba•1h ago
Not OP but one example is that recent VL models are more than sufficient for analyzing your local photo albums/images for creating metadata / descriptions / captions to help better organize your library.
kejaed•1h ago
Any pointers on some local VLMs to start with?
canyon289•58m ago
You could try Gemma4 :D
vunderba•54m ago
The easiest way to get started is probably to use something like Ollama and use the `qwen3-vl:8b` 4‑bit quantized model [1].

It's a good balance between accuracy and memory, though in my experience, it's slower than older model architectures such as Llava. Just be aware Qwen-VL tends to be a bit verbose [2], and you can’t really control that reliably with token limits - it'll just cut off abruptly. You can ask it to be more concise but it can be hit or miss.

What I often end up doing and I admit it's a bit ridiculous is letting Qwen-VL generate its full detailed output, and then passing that to a different LLM to summarize.

- [1] https://ollama.com/library/qwen3-vl:8b

- [2] https://mordenstar.com/other/vlm-xkcd

BoredPositron•1h ago
I use local models for auto complete in simple coding tasks, cli auto complete, formatter, grammarly replacement, translation (it/de/fr -> en), ocr, simple web research, dataset tagging, file sorting, email sorting, validating configs or creating boilerplates of well known tools and much more basically anything that I would have used the old mini models of OpenAI for.
ktimespi•50m ago
For me, receipt scanning and tagging documents and parts of speech in my personal notes. It's a lot of manual labour and I'd like to automate it if possible.
mentalgear•35m ago
Adding to the Q: Any good small open-source model with a high correctness of reading/extracting Tables and/of PDFs with more uncommon layouts.
vunderba•1h ago
Strongly agree. Gemma3:27b and Qwen3-vl:30b-a3b are among my favorite local LLMs and handle the vast majority of translation, classification, and categorization work that I throw at them.
scrlk•1h ago
Comparison of Gemma 4 vs. Qwen 3.5 benchmarks, consolidated from their respective Hugging Face model cards:

    | Model          | MMLUP | GPQA  | LCB   | ELO  | TAU2  | MMMLU | HLE-n | HLE-t |
    |----------------|-------|-------|-------|------|-------|-------|-------|-------|
    | G4 31B         | 85.2% | 84.3% | 80.0% | 2150 | 76.9% | 88.4% | 19.5% | 26.5% |
    | G4 26B A4B     | 82.6% | 82.3% | 77.1% | 1718 | 68.2% | 86.3% |  8.7% | 17.2% |
    | G4 E4B         | 69.4% | 58.6% | 52.0% |  940 | 42.2% | 76.6% |   -   |   -   |
    | G4 E2B         | 60.0% | 43.4% | 44.0% |  633 | 24.5% | 67.4% |   -   |   -   |
    | G3 27B no-T    | 67.6% | 42.4% | 29.1% |  110 | 16.2% | 70.7% |   -   |   -   |
    | GPT-5-mini     | 83.7% | 82.8% | 80.5% | 2160 | 69.8% | 86.2% | 19.4% | 35.8% |
    | GPT-OSS-120B   | 80.8% | 80.1% | 82.7% | 2157 |  --   | 78.2% | 14.9% | 19.0% |
    | Q3-235B-A22B   | 84.4% | 81.1% | 75.1% | 2146 | 58.5% | 83.4% | 18.2% |  --   |
    | Q3.5-122B-A10B | 86.7% | 86.6% | 78.9% | 2100 | 79.5% | 86.7% | 25.3% | 47.5% |
    | Q3.5-27B       | 86.1% | 85.5% | 80.7% | 1899 | 79.0% | 85.9% | 24.3% | 48.5% |
    | Q3.5-35B-A3B   | 85.3% | 84.2% | 74.6% | 2028 | 81.2% | 85.2% | 22.4% | 47.4% |

    MMLUP: MMLU-Pro
    GPQA: GPQA Diamond
    LCB: LiveCodeBench v6
    ELO: Codeforces ELO
    TAU2: TAU2-Bench
    MMMLU: MMMLU
    HLE-n: Humanity's Last Exam (no tools / CoT)
    HLE-t: Humanity's Last Exam (with search / tool)
    no-T: no think
kpw94•1h ago
Wild differences in ELO compared to tfa's graph: https://storage.googleapis.com/gdm-deepmind-com-prod-public/...

(Comparing Q3.5-27B to G4 26B A4B and G4 31B specifically)

I'd assume Q3.5-35B-A3B would performe worse than the Q3.5 deep 27B model, but the cards you pasted above, somehow show that for ELO and TAU2 it's the other way around...

Very impressed by unsloth's team releasing the GGUF so quickly, if that's like the qwen 3.5, I'll wait a few more days in case they make a major update.

Overall great news if it's at parity or slightly better than Qwen 3.5 open weights, hope to see both of these evolve in the sub-32GB-RAM space. Disappointed in Mistral/Ministral being so far behind these US & Chinese models

coder543•47m ago
> Wild differences in ELO compared to tfa's graph

Because those are two different, completely independent Elos... the one you linked is for LMArena, not Codeforces.

nateb2022•36m ago
> Very impressed by unsloth's team releasing the GGUF so quickly, if that's like the qwen 3.5, I'll wait a few more days in case they make a major update.

Same here. I can't wait until mlx-community releases MLX optimized versions of these models as well, but happily running the GGUFs in the meantime!

bachmeier•2m ago
So is there something I can take from that table if I have a 24 GB video card? I'm honestly not sure how to use those numbers.
ceroxylon•1h ago
Even with search grounding, it scored a 2.5/5 on a basic botanical benchmark. It would take much longer for the average human to do a similar write-up, but they would likely do better than 50% hallucination if they had access to a search engine.
WarmWash•57m ago
Even multimodal models are still really bad when it comes to vision. The strength is still definitely language.
wg0•1h ago
Google might not have the best coding models (yet) but they seem to have the most intelligent and knowledgeable models of all especially Gemini 3.1 Pro is something.

One more thing about Google is that they have everything that others do not:

1. Huge data, audio, video, geospatial 2. Tons of expertise. Attention all you need was born there. 3. Libraries that they wrote. 4. Their own data centers and cloud. 4. Most of all, their own hardware TPUs that no one has.

Therefore once the bubble bursts, the only player standing tall and above all would be Google.

chasd00•1h ago
Not sure why you're being downvoted, the other thing Google has is Google. They just have to spend the effort/resources to keep up and wait for everyone else to go bankrupt. At the end of the day I think Google will be the eventual LLM winner. I think this is why Meta isn't really in the race and just releases open weight models, the writing is on the wall. Also, probably why Apple went ahead and signed a deal with Google and not OpenAI or Anthropic.
wg0•1h ago
I don't know why I am downvoted but Google has data, expertise, hardware and deep pockets. This whole LLM thing is invented at Google and machine learning ecosystem libraries come from Google. I don't know how people can be so irrational discounting Google's muscle.

Others have just borrowed data, money, hardware and they would run out of resources for sure.

greenavocado•1h ago
This remains true so long as advertisers give Google money.
bitpush•28m ago
Why wouldnt advertisers give Google money? Are you noticing any shift in trend?
faangguyindia•22m ago
Same can be said for java, yet google own android.
WarmWash•55m ago
The rumor is also that Meta is looking to lease Gemini similar to Apple, as their recent efforts reportedly came up short of expectations.
whimblepop•30m ago
I recently canceled my Google One subscription because getting accurate answers out of Gemini for chat is basically impossible afaict. Whether I enable thinking makes no difference: Gemini always answers me super quickly, rarely actually looks something up, and lies to me. It has a really bad unchecked hallucination problem because it prioritizes speed over accuracy and (astonishingly, to me) is way more hesitant to run web searches than ChatGPT or Claude.

Maybe the model is good but the product is so shitty that I can't perceive its virtues while using it. I would characterize it as pretty much unusable (including as the "Google Assistant" on my phone).

It's extremely frustrating every way that I've used it but it seems like Gemini and Gemma get nothing but praise here.

logicchains•8m ago
Recently I had a pretty basic question about whether there was a Factorio mod for something so decided to ask it to Gemini, it hallucinated not one but two sadly non-existing mods. Even Grok is better at search.
mhitza•10m ago
At the start of last year Gemma2 made the fewest mistakes when I was trying out self-hosted LLMs for language translation. And at the time it had a non open source license.

Really eager to test this version with all the extra capabilities provided.

mudkipdev•1h ago
Can't wait for gemma4-31b-it-claude-opus-4-6-distilled-q4-k-m on huggingface tomorrow
entropicdrifter•31m ago
I'd rather see a distill on the 26B model that uses only 3.8B parameters at inference time. Seems like it will be wildly productive to use for locally-hosted stuff
indrora•13m ago
gemma4-31b-it-claude-opus-4-6-distilled-abliterated-heretic-GGUF-q4-k-m
bertili•1h ago
Qwen: Hold my beer

https://news.ycombinator.com/item?id=47615002

xfalcox•1h ago
Comparing a model you can downloads weights for with an API-only model doesn't make much sense.
regularfry•51m ago
My money's on whatever models qwen does release edging ahead. Probably not by much, but I reckon they'll be better coders just because that's where qwen's edge over gemma has always been. Plus after having seen this land they'll probably tack on a couple of epochs just to be sure.
svachalek•58m ago
The Qwen Plus models should be compared to Gemini, not Gemma.
fooker•1h ago
What's a realistic way to run this locally or a single expensive remote dev machine (in a vm, not through API calls)?
matja•59m ago
I'm running Gemma 4 with the llama.cpp web UI.

https://unsloth.ai/docs/models/gemma-4 > Gemma 4 GGUFs > "Use this model" > llama.cpp > llama-server -hf unsloth/gemma-4-31B-it-GGUF:Q8_0

If you already have llama.cpp you might need to update it to support Gemma 4.

evanbabaallos•1h ago
Impressive
heraldgeezer•1h ago
Gemma vs Gemini?

I am only a casual AI chatbot user, I use what gives me the most and best free limits and versions.

daemonologist•1h ago
Gemma will give you the most, Gemini will give you the best. The former is much smaller and therefore cheaper to run, but less capable.

Although I'm not sure whether Gemma will be available even in aistudio - they took the last one down after people got it to say/do questionable stuff. It's very much intended for self-hosting.

worldsavior•58m ago
Gemma is only 10s of billion parameters, Gemini is 100s.
VadimPR•1h ago
Gemma 3 E4E runs very quick on my Samsung S26, so I am looking forward to trying Gemma 4! It is fantastic to have local alternatives to frontier models in an offline manner.
canyon289•59m ago
Hi all! I work on the Gemma team, one of many as this one was a bigger effort given it was a mainline release. Happy to answer whatever questions I can
wahnfrieden•58m ago
How is the performance for Japanese, voice in particular?
canyon289•26m ago
I dont have the metrics off hand, but I'd say try it and see if you're impressed! What matters at the end of the day is if its useful for your use cases and only you'll be able to assess that!
k3nz0•54m ago
How do you test codeforces ELO?
canyon289•27m ago
On this one I dont know :) I'll ask my friends on the evaluation side of things how they do this
azinman2•53m ago
How do the smaller models differ from what you guys will ultimately ship on Pixel phones?

What's the business case for releasing Gemma and not just focusing on Gemini + cloud only?

canyon289•28m ago
Its hard to say because Pixel comes prepacked with a lot of models, not just ones that that are text output models.

With the caveat that I'm not on the pixel team and I'm not building _all_ the models that are used on, its evident there are many models that support the Android experience, from autocomplete on keyboard to image editing.

https://store.google.com/us/magazine/magic-editor?hl=en-US&p...

abhikul0•51m ago
Thanks for this release! Any reason why 12B variant was skipped this time? Was looking forward for a competitor to Qwen3.5 9B as it allows for a good agentic flow without taking up a whole lotta vram. I guess E4B is taking its place.
mohsen1•50m ago
On LM Studio I'm only seeing models/google/gemma-4-26b-a4b

Where can I download the full model? I have 128GB Mac Studio

gusthema•16m ago
They are all on hugging face
tjwebbnorfolk•43m ago
Will larger-parameter versions be released?
canyon289•38m ago
We are always figuring out what parameter size makes sense.

The decision is always a mix between how good we can make the models from a technical aspect, with how good they need to be to make all of you super excited to use them. And its a bit of a challenge what is an ever changing ecosystem.

I'm personally curious is there a certain parameter size you're looking for?

WarmWash•32m ago
Mainline consumer cards are 16GB, so everyone wants models they can run on their $400 GPU.
NekkoDroid•16m ago
Yea, I've been waiting a while for a model that is ~12-13GB so there is still a bit of extra headroom for all the different things running on the system that for some reason eat VRAM.
NitpickLawyer•27m ago
Jeff Dean apparently didn't get the message that you weren't releasing the 124B Moe :D

Was it too good or not good enough? (blink twice if you can't answer lol)

jimbob45•15m ago
how good they need to be to make all of you super excited to use them

Isn't that more dictated by the competition you're facing from Llama and Qwent?

UncleOxidant•11m ago
Something in the 60B to 80B range would still be approachable for most people running local models and also could give improved results over 31B.

Also, as I understand it the 26B is the MOE and the 31B is dense - why is the larger one dense and the smaller one MOE?

philipkglass•42m ago
Do you have plans to do a follow-up model release with quantization aware training as was done for Gemma 3?

https://developers.googleblog.com/en/gemma-3-quantized-aware...

Having 4 bit QAT versions of the larger models would be great for people who only have 16 or 24 GB of VRAM.

_boffin_•34m ago
What was the main focus when training this model? Besides the ELO score, it's looking like the models (31B / 26B-A4) are underperforming on some of the typical benchmarks by a wide margin. Do you believe there's an issue with the tests or the results are misleading (such as comparative models benchmaxxing)?

Thank you for the release.

logicallee•4m ago
Do any of you use this as a replacement for Claude Code? For example, you might use it with openclaw. I have a 48 GB integrated RAM Mac Mini M4 I currently run Claude Code on, do you think I can replace it with OpenClaw and one of these models?
mwizamwiinga•59m ago
curious how this scales with larger datasets. anyone tried it in production?
chrislattner•51m ago
If you want the fastest open source implementation on Blackwell and AMD MI355, check out Modular's MAX nightly. You can pip install it super fast, check it out here: https://www.modular.com/blog/day-zero-launch-fastest-perform...

-Chris Lattner (yes, affiliated with Modular :-)

nabakin•24m ago
Faster than TensorRT-LLM on Blackwell? Or do you not consider TensorRT-LLM open source because some dependencies are closed source?
simonw•41m ago
I ran these in LM Studio and got unrecognizable pelicans out of the 2B and 4B models and an outstanding pelican out of the 26b-a4b model - I think the best I've seen from a model that runs on my laptop.

https://gist.github.com/simonw/12ae4711288637a722fd6bd4b4b56...

The gemma-4-31b model is completely broken for me - it just spits out "---\n" no matter what prompt I feed it.

wordpad•35m ago
Do you think it's just part of their training set now?
simonw•14m ago
If it's part of their training set why do the 2B and 4B models produce such terrible SVGs?
entropicdrifter•32m ago
Your posting of the pelican benchmark is honestly the biggest reason I check the HackerNews comments on big new model announcements
jckahn•13m ago
All hail the pelican king!
bertili•27m ago
The timing is interesting as Apple supposedly will distill google models in the upcoming Siri update [1]. So maybe Gemma is a lower bound on what we can expect baked into iPhones.

[1] https://news.ycombinator.com/item?id=47520438

whhone•16m ago
The LiteRT-LM CLI (https://ai.google.dev/edge/litert-lm/cli) provides a way to try the Gemma 4 model.

  # with uvx
  uvx litert-lm run \
    --from-huggingface-repo=litert-community/gemma-4-E2B-it-litert-lm \
    gemma-4-E2B-it.litertlm
DeepYogurt•14m ago
maybe a dumb question but what what does the "it" stand for in the 31B-it vs 31B?
bigyabai•10m ago
Instruction Tuned. It indicates that thinking tokens (eg <think> </think>) are not included in training.
virgildotcodes•8m ago
Downloaded through LM Studio on an M1 Max 32GB, 26B A4B Q4_K_M

First message:

https://i.postimg.cc/yNZzmGMM/Screenshot-2026-04-03-at-12-44...

Not sure if I'm doing something wrong?

This more or less reflects my experience with most local models over the last couple years (although admittedly most aren't anywhere near this bad). People keep saying they're useful and yet I can't get them to be consistently useful at all.

solarkraft•2m ago
Wow, just like its larger brother!

I had a similar experience running Qwen 3.5 35b a3b directly through llama.cpp. It would massively overthink every request. Somehow in OpenCode it just worked.

I think it comes down to temperature and such (see daniel‘s post), but I haven’t messed with it enough to be sure.