FW-Gemma-4-26B-A4B-IT

Models available for use with Fireworks on Foundry deliver optimized, best-in-class performance on the Fireworks Inference Cloud. Fireworks on Foundry is a Non-Microsoft Product. The following terms apply to a Customer's use of Fireworks on Foundry: When you use Fireworks on Foundry, data is shared between Microsoft and Fireworks AI, Customer Data will be sent outside of Microsoft systems, Customer Data will not be processed pursuant to any Foundry data residency documentation, and different compliance and data handling rules will apply. See Trust Center - Fireworks AI for details. Customers are responsible for evaluating whether data sharing between Microsoft and Fireworks is appropriate for their organization's compliance requirements.

About this model

Gemma 4 26B A4B IT is an instruction-tuned multimodal Mixture-of-Experts model from Google DeepMind with 25.2B total parameters and 3.8B active parameters. It supports text and image input with text output, includes configurable thinking modes for reasoning, and uses a ~550M-parameter vision encoder with variable aspect ratio and resolution support. This variant has 30 layers, a 1,024-token sliding window, and a 256K-token context window, and the Gemma 4 family maintains multilingual support in over 140 languages.

Key model capabilities

• Text and image input with text output
• Configurable thinking modes for reasoning
• 256K-token context window
• Multilingual support in over 140 languages
• Variable aspect ratio and resolution support for images