Gemma 2 2B is here! Fantastic performance for size, it's great for research and applications. I am very proud of the progress our team made over the last few months!

Gemma 2 - 2B is live! The results are incredible for a small model -- better on LMSys Chatbot Arena than all GPT-3.5 models and besting Llama-2-70B-chat, which was very recently the biggest and best model! The pace of change is amazing.

Last but not least, our new Gemma 2 2B. Our smallest model which performs amazingly well both on pre-training and post-training benchmarks! Looking forward to seeing what people will build with it, in particular for on-device applications!

Gemma 2 2B is released today! Only takes up ~2.5 Gb on int8 quant, and scores better than GPT 3.5/ Mixtral 8x7B on Chatbot Arena -- great performance for its size :) Super happy with the amazing progress from the team!

Did I mention that Gemma 2 2B performs better on LMSYS than Mixtral 8x7B, GPT 3.5 and Llama 2 70B? 🤯🤯

🛡️ ShieldGemma is now available! Explore how the series of safety classifiers empower devs and researchers to build AI with safety in mind → goo.gle/46rOJ2i ⚡️ SOTA performance 🌟 Open classifiers 💫 Flexible model sizes for various use cases 👀 And so much more!

I guess it's official now: I'm working on the Open Model Sam Altman announced earlier this month. Thanks all for joining our listening session: amazing feedback and eager to build great things together!

Doing some research on the use cases for guardrails/classifiers : please share with me how you are using it and what's most helpful: flexibility (i can adapt to my policies)? Latency? Performance? Other? cc/ Sayash Kapoor Joseph Spisak Victor Storchan Dana Kurniawan - tag the right folks!

we are going to take a little more time with our open-weights model, i.e. expect it later this summer but not june. our research team did something unexpected and quite amazing and we think it will be very very worth the wait, but needs a bit longer.

we planned to launch our open-weight model next week. we are delaying it; we need time to run additional safety tests and review high-risk areas. we are not yet sure how long it will take us. while we trust the community will build great things with this model, once weights are

Our open models are here. Both of them. openai.com/open-models

gpt-oss is out! we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!) (and a smaller one that runs on a phone). super proud of the team; big triumph of technology.

gpt-oss is a big deal; it is a state-of-the-art open-weights reasoning model, with strong real-world performance comparable to o4-mini, that you can run locally on your own computer (or phone with the smaller size). We believe this is the best and most usable open model in the

We're out! Try it and tell us what you build !

Thanks clem 🤗 for the amazing collaboration here! ➡️ Go check out HF: huggingface.co/openai