Currently listed through these providers:
Model details
Gemma 4 26B A4B
Gemma 4 26B A4B is an instruction-tuned member of Google DeepMind's Gemma 4 family of open-weight releases. It sits alongside four sibling sizes that span a range of deployment targets, from phones and laptops up to server-class hardware, giving the family unusual flexibility for a single generation. The broader Gemma 4 line is described as multimodal, accepting text and image input while producing text output, and is built to handle extended contexts of up to 256K tokens across more than 140 languages.
The 26B A4B designation marks this variant as a mixture-of-experts configuration, paired with the family's dense options, which reflects a deliberate split between pure-capacity and routed-capacity designs for different efficiency goals. Across Gemma 4, reasoning is a first-class capability with configurable thinking modes, and the family is positioned for text generation, coding, and reasoning workloads. Open weights and an Apache 2.0 license make the 26B A4B instruction-tuned variant well suited to teams that want to self-host, fine-tune, or experiment with a mid-to-large MoE while staying inside an officially supported Google DeepMind lineage.
Quick Info
Powered by- Provider
- NanoGPT
- Model key
- google/gemma-4-26b-a4b-it
- Release date
- Apr 2, 2026
- Last updated
- Apr 2, 2026
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $0.12
- Output token cost
- $0.38
Limits
- Input tokens
- 262,144 tokens
- Output tokens
- 131,072 tokens
- Context window
- 262,144 tokens
Transparent token rates
Compare Gemma 4 26B A4B pricing
Rates are shown per one million tokens. Combined means one million input plus one million output tokens.
Latest news about Gemma 4 26B A4B
No articles yet. Fetch the latest news to show it here.