Q4_K_M
The conversion repository lists gemma-4-26B-A4B-it-Q4_K_M.gguf at 16.8 GB; its rounded decimal file size is converted to GiB for planning. This does not include context or runtime memory.
- Bits per weight
- 4.5
- Model size
- 15.65 GiB
Model catalog
Gemma 4 26B-A4B IT
Google's multimodal MoE has 25.2B total parameters and 3.8B active parameters per token. Weight planning uses total parameters; the separate vision projector is excluded.
Model overview
Available quantizations
Lower-bit quantizations generally use less memory. The compatibility calculator uses these candidates and its transparent approximate memory assumptions; it does not make a performance guarantee.
Compatibility notes
Estimates use the cataloged model-weight size (or a parameter-based estimate), apply weight overhead, and include standard runtime overhead. They do not include context/KV-cache memory. Real requirements vary with runtime, drivers, and other system use. A model page cannot determine compatibility without your hardware profile; use the calculator for that assessment.
Google Gemma 4 26B-A4B IT model card; verified value, last verified 2026-09-29. Google lists 25.2B total parameters, 3.8B active parameters, Apache-2.0 licensing, and 256K context. Weight planning uses the total count. The GGUF repository lists a separate vision projector file. View source
Multimodal memory scope: This estimate excludes the separate vision/projector file and its runtime memory. ggml-org Gemma 4 26B-A4B IT GGUF repository