Gemma 4... intelligence for everyone on device!

Gemma 4 has been installed like 200,000,000 times : )

Damn! ok, I stand corrected.

@OfficialLoganK its just distribution advatage man nobody uses this seriously in the age of glm 5.2

@OfficialLoganK Oh yeah, GLM 5.2 absolutely crushes Gemma on benchmarks.

Gemma models are amazing. Thank you so much for open sourcing them. We need mooaaaar

A local default 💪👀

ngl these open models are some of the best things you've released recently would be cool if you guys niched down into this direction

OK, can someone explain why Gemma 4 is still relevant in a world where GLM 5.2 exists ?

nice work on gemma models. more love for gemini models is also appreciated

We finetuned Gemma 4 to make WispFlow like dictation run on your macOS. Open source and free! Try us out and save your $15 to get Google pro subscription ;)

Gemma models 🗿🗿 Gemini models 🗿🗿 when???

Where

Gemini models have been completely unusable the past two days why did you nuke them

Gemma 4 ❤️

Gemma4 is love ❤️

I can agree and they serve even better as dedicated subagents

Literally, anything but Gemini 3.5 Pro

Gemma 4 is the unsung hero. I helped so many ppl run it locally. The best story though gotta be my 10 years old PC with a 980Ti that is now serving the 26B at 35 tps as a donation for someone 😂

been testing them all out (and using lm studio link to access all from iphone)

loving the gemma 4 family. impressive as standalone models, but even more useful as dedicated subagent models

amazing.

Nice work, keep the good work @Google

eko is actually powered by gemma 4👇

Gemma is the new Qwen

3.5 Pro not today

Great model for scientific applications on device.

Local AI models are the only way

🔥🔥

🚀 this !!

Gemma 4 are amazing. 31b is my favourite but the tiny ones are great on mobile and for small tasks.

I love Gemma 4!

@grok can it run on my mac mini 16gb?

on device intelligence for all

Google are doing an incredible job with Gemma. Crazy that this runs really well on a phone. 4 years ago this would have seemed impossible

This is the way

Gemma 4 on-device is actually huge 🔥 Finally frontier-level intelligence that works offline.

Congrats Logan.

@LyalinDotCom This is awesome, good work!