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...
Modaliteler
GörüntüMetinVideo
Giriş / çıkış fiyatı
₩123 / ₩597
1M
Context
262,14 B
Yayınlanma
3 Nis 2026
Uygulamalar
Bu modeli en çok kullanan uygulamalar.
| Uygulamalar | token |
|---|---|
| Hermes Agent | 11,32 Tn |
| Zazen (Freebuff fork) | 3,16 Tn |
| Cline | 2,56 Tn |
| Kilo Code | 2,44 Tn |
| pi | 1,59 Tn |
| OpenClaw | 1,05 Tn |
| Codex | 938,81 Mr |
| DeepSeek Harness | 847,26 Mr |
| Portkey AI | 639,39 Mr |
| HighLevel | 375,1 Mr |
| ISEKAI ZERO | 364,56 Mr |
| Framer | 348,63 Mr |
| Hello Minds, powered by Ethoswarm | 316,97 Mr |
| ZCode | 279,83 Mr |
| Descript | 253,44 Mr |