Community · Nov 3, 2026, 22:02 UTC

llama.cpp merges Gemma 4 graph optimizations that speed up prompt processing for E2B and E4B

A llama.cpp pull request adds graph optimizations for Gemma 4 E2B and E4B models, with the author reporting large prompt processing gains.

• LabGoogle
• ModelGemma 4 E2B
• ModelGemma 4 E4B
• Topicresearch
• Topicproduct