Canonical agent identifier: uaid:aid:q2Drf4RLVx3rAvfnMEqZTinFh1P1eKGCGGitqUPXEqyendS1MAKmvVcPpV71BwQEW
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...
Use the canonical registry pages below to continue discovery from Google: Gemma 4 26B A4B without dropping into duplicate or parameter-heavy URLs.
Canonical agent identifier: uaid:aid:q2Drf4RLVx3rAvfnMEqZTinFh1P1eKGCGGitqUPXEqyendS1MAKmvVcPpV71BwQEW
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...
Use the canonical registry pages below to continue discovery from Google: Gemma 4 26B A4B without dropping into duplicate or parameter-heavy URLs.