Gemma 4 26B A4B
Active Parameters
25.2B
Context Length
256K
Modality
Multimodal
Architecture
Mixture of Experts (MoE)
License
Apache 2.0
Release Date
2 Apr 2026
Knowledge Cutoff
-
System Requirements
VRAM requirements for different quantization methods and context sizes
1,024 tokens
Consumer
3x RTX 4090
24GB VRAM
Datacenter
1x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
256,000 tokens
Consumer
6x RTX 4090
24GB VRAM
Datacenter
2x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
Architecture Diagram
Evaluation Benchmarks
Rank
#53
| Benchmark | Score | Rank |
|---|---|---|
Web Development | 1362 | 66 |
General Text | 1438 | 66 |
Rankings
Overall Rank
#53
Coding Rank
#64
About Gemma 4 26B A4B
Gemma 4 26B A4B is a Mixture-of-Experts model with 25.2B total parameters but only 3.8B active per inference, achieving the speed of a 4B model with near-31B performance. Features 128 experts (8 active) with 256K context window, supporting text and image input. Optimized for fast inference on consumer GPUs while delivering frontier-level reasoning and coding capabilities.
Technical Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
16
Key-Value Heads
8
Attention Head Dimension
256
Position Embedding
ROPE
RoPE Theta
10,000
Sliding Window Attention
Yes
Sliding Window Size
1,024
Sliding Window Ratio
83.3%
Linear Attention
No
Linear Attention Ratio
-
Normalization
RMS Normalization
Activation Function
GELU
Dimensions
Hidden Dimension Size
2,112
Number of Layers
30
FFN Intermediate Size (Dense)
704
Multi-Token Prediction Heads
-
Tokenizer
Vocabulary Size
262,144
Mixture of Experts
Total Expert Parameters
3.8B
Number of Experts
128
Active Experts
8
Shared Experts
-
FFN Intermediate Size (per Expert)
704
Dense Layers Before MoE
-
About Gemma 4
Gemma 4 is Google DeepMind's most advanced open model family, built from Gemini 3 research and technology. Featuring both Dense and Mixture-of-Experts (MoE) architectures, these multimodal models handle text, images, and audio (on smaller variants), with context windows up to 256K tokens. Designed for frontier-level performance across reasoning, coding, and agentic workflows, Gemma 4 delivers unprecedented intelligence-per-parameter from mobile devices to enterprise servers. Released under Apache 2.0 license.