Gemma 4 12B Instruct | BIG-Bench Extra Hard Original — micro average

Google (DeepMind)'s Gemma 4 12B Instruct evaluation results on BIG-Bench Extra Hard Original — micro average

Score: 53.0%

Reasoning level: thinking mode

Source

Gemma 4 12B Instruct (Registry No. 35005)

Google (DeepMind)

BIG-Bench Extra Hard Original — micro average