Weekend model updates!
Gemma 4 is available for encrypted inference to all users. It supports image upload and reasoning.
DeepSeek R1 has been retired. Thank you for your service. We are keeping watch of new DeepSeek developments.
We are exploring Gemma 4 as one of the default models, probably replacing GPT-OSS-120b as the "Quick" model.
Try it out yourself and let me know which one you like better for quick responses: Gemma 4 vs GPT-OSS
View quoted note โ