Gemma 2 27B

Google's 27-billion-parameter open-weight text model with an 8K token context window.

Strengths

Balances model capacity with inference speed, offering reasonable performance across general text tasks without requiring cutting-edge hardware.

Best for

Self-hosted deployments and local development where you need a mid-sized model that can run on commodity GPUs.

Limitations

The 8K context window is significantly smaller than later Gemma releases (Gemma 3 27B raised this to 110K), and it predates Gemma 4's extended 262K context and improved capabilities.