STORY · MODELLER_
ZAYA1-8B matches DeepSeek-R1 in mathematics with under 1 billion active parameters
A new open model called ZAYA1-8B performs at the level of DeepSeek-R1 in mathematics and coding tasks while using fewer than 1 billion active parameters through efficient sparse activation. This makes it significantly more resource-efficient to run than larger competitors.
WHY IT MATTERS
Demonstrates that you don't need enormous models to achieve high performance on specialized tasks – sparse activation is becoming increasingly important for practical AI use. Lowers the barrier for running powerful reasoning locally or on smaller devices.
SOURCES
MACHINE-GENERATED SUMMARY This summary is written by machine from the sources below. We sort and explain — but we are a way into the field, not the final word. Check the source when something matters to you.