3 ms·
What do you mean it's on ollama and requires h100? As a proprietary google model, it runs on their own hardware, not nvidia.
by grim_io 8mo ago
What do you mean it's on ollama and requires h100? As a proprietary google model, it runs on their own hardware, not nvidia.
- KaiserPro 8mo agosorry A lack of context: https://ollama.com/library/gemini-3-pro-preview https://ollama.com/library/gemini-3-pro-preview You can run it on your own infra. Anthropic and openAI are running off nvidia, so are meta(well supposedly they had custom silicon, I'm not sure if its capable of running big models) and mistral. however if google really are running their own inference hardware, then that means the cost is different (developing silicon is not cheap...) as you say.
- zozbot234 8mo agoThat's a cloud-linked model. It's about using ollama as an API client (for ease of compatibility with other uses, including local), not running that model on local infra. Google does release open models (called Gemma) but they're not nearly as capable.
- simonw 8mo agoYou can't run Gemini 3 Pro Preview on your own infrastructure. Ollama sell access to cloud models these days. It's a little weird and confusing.
- KaiserPro 8mo agoAhh fuck, thanks for pointing that out. I did think its a bit weird that they had open-weighted it