@_philschmid: Weights: https://huggingface.co/collections/google/gemma-4-qat-q4-0… Blog: https://blog.google/innovation-and-ai/techno…
Summary
Google released Gemma 4 models with quantization-aware training (QAT) at Q4_0 precision on Hugging Face, offering efficient variants from 5B to 33B parameters.
View Cached Full Text
Cached at: 06/08/26, 03:22 PM
Weights: https://huggingface.co/collections/google/gemma-4-qat-q4-0… Blog: https://blog.google/innovation-and-ai/technology/developers-tools/quantization-aware-training-gemma-4/…
Gemma 4 QAT Q4_0 - a google Collection
Source: https://huggingface.co/collections/google/gemma-4-qat-q4-0
- —
#### google/gemma-4-E2B-it-qat-q4_0-unquantized Any-to-Any• 5B• Updated3 days ago • 1.92k • 10 - — #### google/gemma-4-E4B-it-qat-q4_0-unquantized Any-to-Any• 8B• Updated3 days ago • 1.39k • 6
- — #### google/gemma-4-12B-it-qat-q4_0-unquantized Any-to-Any• 12B• Updated3 days ago • 4.52k • 32
- — #### google/gemma-4-26B-A4B-it-qat-q4_0-unquantized Image-Text-to-Text• 27B• Updated3 days ago • 1.61k • 16
- — #### google/gemma-4-31B-it-qat-q4_0-unquantized Image-Text-to-Text• 33B• Updated3 days ago • 1.81k • 13
- — #### google/gemma-4-E2B-it-qat-q4_0-unquantized-assistant Any-to-Any• 78M• Updated3 days ago • 158 • 4
- — #### google/gemma-4-E4B-it-qat-q4_0-unquantized-assistant Any-to-Any• 78.8M• Updated3 days ago • 221 • 3
- — #### google/gemma-4-12B-it-qat-q4_0-unquantized-assistant Any-to-Any• 0.4B• Updated3 days ago • 1.12k • 13
- — #### google/gemma-4-26B-A4B-it-qat-q4_0-unquantized-assistant Image-Text-to-Text• 0.4B• Updated3 days ago • 399 • 6
- — #### google/gemma-4-31B-it-qat-q4_0-unquantized-assistant Image-Text-to-Text• 0.5B• Updated3 days ago • 1.13k • 12
- — #### google/gemma-4-E2B-it-qat-q4_0-gguf Any-to-Any• 5B• Updated3 days ago • 9.57k • 26
- — #### google/gemma-4-E4B-it-qat-q4_0-gguf Any-to-Any• 7B• Updated2 days ago • 12.8k • 27
- — #### google/gemma-4-12B-it-qat-q4_0-gguf Any-to-Any• 12B• Updated3 days ago • 52.4k • 85
- — #### google/gemma-4-26B-A4B-it-qat-q4_0-gguf Image-Text-to-Text• 25B• Updated3 days ago • 18.6k • 42
- — #### google/gemma-4-31B-it-qat-q4_0-gguf Image-Text-to-Text• 31B• Updated3 days ago • 12.9k • 49
- — #### google/gemma-4-E2B-it-qat-w4a16-ct Any-to-Any• 6B• Updated3 days ago • 2.81k • 3
- — #### google/gemma-4-E4B-it-qat-w4a16-ct Any-to-Any• 9B• Updated3 days ago • 11.1k • 3
- — #### google/gemma-4-12B-it-qat-w4a16-ct Any-to-Any• 13B• Updated3 days ago • 152k • 16
- — #### google/gemma-4-31B-it-qat-w4a16-ct Image-Text-to-Text• 34B• Updated3 days ago • 22.6k • 16
Similar Articles
google/gemma-4-12B-it-qat-q4_0-gguf
Google DeepMind releases Gemma 4 models optimized with Quantization-Aware Training (QAT) in multiple formats including GGUF, enabling high quality with reduced memory requirements.
Gemma 4 QAT models: Optimizing compression for mobile and laptop efficiency
Google releases Gemma 4 models optimized with Quantization-Aware Training (QAT) to improve efficiency for mobile and laptop deployment, reducing memory footprint to 1GB for the E2B model while preserving quality.
@TheAhmadOsman: Great news Google just released the QAT (4bit) of their Gemma 4 model series including the 31B Dense and the 26B MoE An…
Google released QAT (4-bit) versions of their Gemma 4 model series, including the 31B Dense and 26B MoE models, furthering open-source AI.
Google's quantization aware trained Gemma checkpoints enabling mobile device inference just dropped on HF
Google released quantization-aware trained Gemma 4 checkpoints on HuggingFace, optimized for mobile device inference and available in QAT Mobile and Q4_0 variants.
google/gemma-4-26B-A4B-it
Google DeepMind releases Gemma 4, a family of open-weight multimodal models ranging from 2.3B to 31B parameters with support for text, image, video, and audio inputs. The models feature 256K context windows, MoE and dense architectures, enhanced reasoning capabilities, and are optimized for deployment across devices from mobile to servers.