Gemma 2 (9B IT)
Gemma · ~9B parameters · 4B–12B (mid-size) · Last reviewed 2026-03-20
Why we use it
Balanced mid-size instruct model for multilingual pilots where teams already standardize on Google AI tooling.
License summary
Gemma terms of use; redistribution and acceptable-use rules differ from Apache - read the current Google policy.
Typical deployment profiles
- VPC, single GPU class
- VPC, multi-GPU
Focus tags
- General
- Multilingual
Typical use cases
- Multilingual helpdesks
- Teacher checkpoints
- VPC chat