Gemma 3 12B Specs, Benchmarks & Downloads
Gemma 3 12B
Google mid-size multimodal model
Status
active
Parameters
12.00
Context Window
128000
Modality
Vision
Release Date
Mar 12, 2025
About
Explore Gemma 3 12B, a vision AI model developed by Google. View specifications, context window, architecture, license, supported platforms, features, benchmarks, official downloads, and related articles.
Gemma 3 12B provides stronger reasoning and image understanding capabilities.
Family
GemmaDeveloper
GoogleArchitecture
Vision Language ModelLicense
Gemma LicenseHardware Requirements
Minimum:
Use Case: Multimodal inference
GPU Minimum: RTX 3060
GPU VRAM: 12GB
System RAM: 32GB
Storage: 50GB
Operating System: Linux
12B vision model
Recommended:
Use Case: Vision inference
GPU Minimum: RTX 4090
GPU VRAM: 24GB
System RAM: 64GB
Storage: 80GB
Operating System: Linux
Recommended multimodal deployment
High Performance:
Use Case: Production vision inference
GPU Minimum: 2x RTX 4090
GPU VRAM: 48GB
System RAM: 128GB
Storage: 150GB
Operating System: Linux
High throughput vision
Model Deployment Readiness
Model Ecosystem Support
18%
Platforms
33%
Features
1%
Downloads
20%
Platform Compatibility
33%
Supported across 4 of 12 platforms
Context Window
Industry-standard context window comparison
128,000
tokens
125K
4K
32K
128K
1M
2M
Database Rank
#17
Percentile
Top 84%
Capacity
High
Benchmark Performance
Evaluation profile across publicly reported benchmarks
|
MMMU
|
Vision | 59.0 % | 2025 |
|
MMLU
|
Knowledge | 74.5 % | 2025 |
|
GPQA
|
Reasoning | 40.9 % | 2025 |
|
HumanEval
|
Coding | 85.4 % | 2025 |
|
LiveCodeBench
|
Coding | 24.6 % | 2025 |
|
GSM8K
|
Mathematics | 94.4 % | 2025 |
|
IFEval
|
Instruction Following | 88.9 % | 2025 |
|
MMLU
|
Knowledge | 70.0 % | 2025 |
Highest Benchmark
GSM8K
94.4%
Benchmarks Available
8
Latest Evaluation
Mar 2025
