Side-by-side comparison of Gemma 4 (Google · USA) and Gemma 4 26B A4B Instruct (Google DeepMind · United States) for self-hosted deployment of the open-weight model. Gemma 4 is rated EU-ready; Gemma 4 26B A4B Instruct is conditional. They part ways on commercial use: Gemma 4 is "Yes", Gemma 4 26B A4B Instruct is "Unrestricted".
| Field | ||
|---|---|---|
| Summary | ||
| Verdict | EU-ready Major licence shift from Gemma 2/3: Apache 2.0 across the family. 140+ languages, multimodal (text/image/audio/video on small sizes), 128K-256K context. Strong permissive default for EU deployments that need robust multilingual support. | Conditional Based on published licence terms, Gemma 4 26B A4B ships under pure Apache 2.0 with no prohibited-use carve-outs — a departure from prior Gemma generations. The sparse-MoE architecture (25.2B total / 3.8B active) puts it in an ambiguous zone for EU AI Act GPAI systemic-risk classification, and US origin plus image-input support add transparency obligations that deployers should document. |
| Last reviewed | 2026-04-15 | 2026-04-17 |
| Open-weight | ||
| Licence | Apache 2.0 | Apache 2.0 |
| Commercial use | Yes | Unrestricted |
| Training data | Partial | Domain-level summary |
| Origin | USA | United States |
| Performance & pricing? | ||
| Quality index | 32/100 | 27/100 |
| Speed | — | — |
| Blended price | — | — |
| Context window | — | — |
| Evidence | ||
| Sources | ||
No overlapping sources between the two entries.