codegemma:2b

Upstream reference: CodeGemma 2B pretrained

Publisher: Google

Google documents the 2B CodeGemma variant as a pretrained code model for completion, distinct from the 7B instruction-tuned conversational model.

Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.

Published upstream specifications

Published size label

2 billion parameters Source

Tuning stage

Pretrained, not the 7B instruction-tuned variant Source

Documented completion type

Prefix/suffix and fill-in-the-middle Source

Publisher-described uses

• IDE-style code completion and infilling. Source

Reference limitations

• Chat/instruction behavior from the 7B IT variant must not be assigned to this 2B pretrained checkpoint. Source

A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.

Publisher sources

• CodeGemma 2B pretrained official documentation — Google. Checked 2026-11-02.

Links

Recorded references do not establish downloadable weights, licensing or current availability.

• Google AI / Google DeepMind — official site
• Ollama
• Provider page
• Hugging Face
• Official model announcement / documentation

Where codegemma:2b comes from

Named in 2 observation pages

• A consumer GPU recorded higher throughput than a 48-core Xeon · V1
• NVIDIA GTX 1070 — 344 models profiled · V1