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.

ProviderContextQuantizationMax outputPriorityState
OpenRouter262,14 rb50unknown