Google: Gemma 4 26B A4B
google/gemma-4-26b-a4b-it
Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...
Modalities
GambarTeksVideo
Harga in / out
₩123 / ₩597
1M
Context
262,14 rb
Dirilis
3 Apr 2026
Providers
Beberapa tempat menghosting model yang sama. Permintaan disebarkan berdasarkan prioritas dan kesehatan, dan local GPU selalu didahulukan jika tersedia.
| Provider | Context | Quantization | Max output | Priority | State |
|---|---|---|---|---|---|
| OpenRouter | 262,14 rb | — | — | 50 | unknown |