Model
Inference & ServingCoding
CodeGemma
1Google's 2Gemma 3variant 4tuned 5for 6code
the six Ws · specification
W1
Who
Google DeepMind released CodeGemma as a code-specialized variant of Gemma.
W2
What
A family of 2B and 7B models fine-tuned for code completion, generation, and chat.
W3
Where
Published on Hugging Face and Kaggle under the Google organization.
W4
When
CodeGemma was released in April 2024.
W5
Why
It extended the lightweight Gemma family into developer-focused code tasks.
W6
With
Runs via transformers, Ollama, or llama.cpp, built on the Gemma base architecture.
W7
Watch
Open-weight release by Google under the Gemma license, which permits broad use with some usage restrictions.
alternatives
works with