Google releases Gemma 4 open models • Hacker News Board

Re: Google releases Gemma 4 open models

#411

Post by kristjansson » Fri, Apr 03, 2026, 7:03 AM UTC

Wowfunhappy wrote: post #380

car wrote: post #368

Earlier quoted context omitted.

Yes, they are listed on huggingface. The instruction trained models have an 'it' in their name. https://huggingface.co/collections/unsloth/gemma-4 Edit: Sorry, I'm not sure if this is a quant, but it says 'finetuned' from the Google Gemma 4 parent snapshot. It's the same size as the UD 8-bit quant though.

Only the 'it' models seem to have quants. I was really hoping to try a base model.

Basic quantization is easy if you have enough RAM (not VRAM) to load the weights.

Re: Google releases Gemma 4 open models

#412

Post by petu » Fri, Apr 03, 2026, 7:05 AM UTC

swalsh wrote: post #341

redman25 wrote: post #280

Earlier quoted context omitted.

Exactly, compare MoE with MoE and dense with dense otherwise it's apples and oranges.

Its coding to coding. I could care less how the model is architected, i only care how it performs in a real world scenario.

If you don't care about how it's architectured, why you care about size? Compare it to Q3.5 397B-A17B.

Just like smaller size models are speed / cost optimization, so is MoE.

G4 26B-A4B goes 150 t/s on 4090/5090, 80 t/s on M5 Max. Q3.5 35B-A3B is comparably fast. They are flash-lite/nano class models.

G4 31B despite small increase in total parameter count is over 5 times slower. Q3.5 27B is comparably slow. They are approximating flash/mini class models (I believe sizes of proprietary models in this class are closer to Q3.5 122B-A10B or Llama 4 Scout 109B-A17B).

Re: Google releases Gemma 4 open models

#413

Post by ggnore7452 » Fri, Apr 03, 2026, 7:13 AM UTC

too bad that only the smaller on-device models support native audio input.

Re: Google releases Gemma 4 open models

#414

Post by sixhobbits » Fri, Apr 03, 2026, 7:20 AM UTC

danielhanchen wrote: post #2

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 " ". " thought\n" is also used for the thinking trace!

Thanks for this, I gave this guide to my Claude and he oneshot the unsloth and gemma4 set up on the old macbook he runs on. It's way faster than I expected, haven't tried out local models for a few generations but will be very nice when they become useful

Re: Google releases Gemma 4 open models

#415

Post by logicallee » Fri, Apr 03, 2026, 7:29 AM UTC

downrightmike wrote: post #287

logicallee wrote: post #117

Earlier quoted context omitted.

Do any of you use this as a replacement for Claude Code? For example, you might use it with openclaw. I have a 24 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?

Did you try it?

yes, I've now I tried both the 20 GB version (gemma4:31b) which is the largest on the page[1], and the ~10 GB version (gemma4:e4b). The 20 GB version was rather slow even when fully loaded and with some RAM still left free, and the 10 GB version was speedy. I installed openclaw but couldn't get it to act as an agent the way Claude Code does. If you'd like to see a video of how both of them perform with almost nothing else running, on a Mac Mini M4 with 24 GB of RAM, you can see one here (I just recorded it):[2]

[1] https://ollama.com/library/gemma4

[2] https://www.youtube.com/live/G5OVcKO70ns

Re: Google releases Gemma 4 open models

#416

Post by petu » Fri, Apr 03, 2026, 7:35 AM UTC

stavros wrote: post #344

g947o wrote: post #335

Earlier quoted context omitted.

*8 months That doesn't make parent's claim true or even relevant. And OpenAI could release an open model tomorrow. Nobody knows.

Anybody could release an open model tomorrow. Google is the only US based lab releasing open weights models. OpenAI released one once, which might or might not count as "releasing", depending on your definition

> OpenAI released one once

they've released gpt-oss-safeguard in October

I hope / think they are going to release more, just going for one big release a year like Gemma (if we talk strictly about general chat model -- Gemma 3 was March 2025)

Re: Google releases Gemma 4 open models

#417

Post by alecthomas » Fri, Apr 03, 2026, 7:42 AM UTC

nl wrote: post #375

Gemma-4-E4B-it scored 15/25 on my https://sql-benchmark.nicklothian.com/#all-data (agentic SQL generation). The naming is a bit odd - E4B is "4.5B effective, 8B with embeddings", so despite the name it is probably best compared with the 8B/9B class models and is competitive with them. Qwen3.5-9B also scores 15/25 in thinking mode for example. The best 9B model I've found is Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distil…

Oh this page is great! I just released AIM [1] which is a tool that generates verified SQL migrations using LLMs, and I tested a bunch of models manually. I think I'll just link to your page too!

[1] https://github.com/alecthomas/aim