Gemma
Google DeepMind's family of open AI models, small enough to run on your own computer.
Recently discussed in
What creators say
"Gemma. This is Google's open model. This one runs remarkably small. There's actually a version that fits in 16 GB of RAM. And that one can fit on your phone. It's beautiful, clean writing. The fact that Google gives this away for free is actually ..."
More from Greg Isenberg →+1 more mention from Greg
"Google has Gemini, but they've also got Gemma, and Gemma's their open source."
"They asked Gemma, that's the model running on board, to write a haiku about being in space."
More from This Week in Startups →+1 more mention from This
"I'm just worried that the models from Nvidia's Neatron family, the models from Google's Gemma family, aren't sufficient to replace what we have fro..."
"Google has something called Gemma 4, which is also very powerful, a smaller model. You can download these things."
More from Eric Schmidt →"We have the incredible models — the original Llama models of course, but also the Gemma models from DeepMind. These are great choices for local."
More from Jeffrey Morgan →"There are like Gemma for example — these companies like to support open source because it actually helps build the ecosystem around the things that they're building."
More from Chelsea Finn →
"Demis's quasi public apology for why Gemini itself wasn't open-weight — the claim was we would open-weight it, open source it if we had the compute, but we don't have the compute. So instead we only have the compute available for Gemma."
More from Emad Mostaque →"Google has Gemma as well. Gemma, by the way, is pretty good in terms of efficiency. If you look at Arena, you'll see that on the Pareto curves of performance versus cost, Gemma's on there."
More from Anastasios Angelopoulos →"We did about 20 plus different state-of-the-art local models across Gemma, GPT OSS, Qwen, IBM Granite, all in the range of one to 200 billion parameters."
More from Jon Saad-Falcon →
"We've got Google Gemma 4. But that changed last night with some breaking news."
More from Peter Diamandis →
"You can still run local models on them. There's options out there like Gemma 4, like a few other really small ones. They're not going to be anything close to frontier, but you can do small things like embedding."
More from Alex Finn →
"One of the most downloaded models on Ollama is a US model — it's the Gemma models. And this is like a super amazing team at DeepMind that's putting them out. The new ones are agent-ready. They can run coding agent loops. They can accomplish much h..."
More from John Coogan →"We are also pushing hard on a suite of open source models called Gemma which we're determined to make best-in-class for their sizes."
More from Demis Hassabis →"I looked at the benchmark eval for it. It looks — I mean it's stronger than Gemma 4, but on the other hand that's not saying very much because Gemma 4 isn't that strong."
Keep exploring
Topics
About this data
- ✓48,000+ mentions extracted from real podcast transcripts
- ✓Every quote links to the source episode
- ✓Sponsored mentions are always labeled
- ✓Critical takes are shown, never filtered out