Overview
CodeGemma continued training Gemma on about 500 billion tokens of mostly code, adding fill-in-the-middle support for editor completion.
Highlights
- 2B and 7B sizes
- Fill-in-the-middle completion
- Instruction-tuned 7B variant
- Runs on laptops
Typical uses
Local code completion in editors and lightweight coding chat.
Things to consider
- Older than current code models; short 8K context.
- Download sizes refer to the default quantized versions in the Ollama library; full-precision weights are larger.
Specifications are taken from the model card on Hugging Face, the Ollama library and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.