Gemma-3-1B

Gemma 3 1B — a compact language model developed by Google DeepMind, achieving an impressive balance between size and capabilities. With just 1 billion parameters and a context window of 32,000 tokens, the model is highly efficient and capable of running on devices with limited resources. Thanks to its architecture, Gemma 3 1B optimizes memory usage, making it an ideal choice for embedded systems and mobile applications.

It's important to note that the 1B version is a text-only model and does not support image processing, unlike the larger variants in the Gemma 3 series (4B, 12B, and 27B), which offer multimodal capabilities.

Additionally, the model has limited language support and is primarily optimized for tasks in English.

Gemma 3 1B is available with open weights, making it easy to fine-tune and adapt to specific use cases. It comes in multiple quantization levels, ranging from 32-bit down to 4-bit, providing added flexibility when deploying on different hardware platforms.

Announce Date: 12.03.2025
Parameters: 1B
Context: 32K
Layers: 26, using full attention: 4
Attention Type: Sliding Window Attention
VRAM requirements: 3.3 GB using 4 bits quantization
Developer: Google DeepMind
Transformers Version: 4.50.0.dev0
Ollama Version: 0.6
License: gemma

Use our pre-built public endpoints for free to test inference and explore Gemma-3-1B capabilities. You can obtain an API access token on the token management page after registration and verification.

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

• maximize endpoint performance,
• enable full context for long sequences,
• ensure top-tier security for data processing in an isolated, dedicated environment,
• use custom weights, such as fine-tuned models or LoRA adapters.

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.