ZAYA1-8B: An 8B Moe Model with 760M Active Params Matching DeepSeek-R1 on Math

Researchers have released an open-source 8B Moe model called ZAYA1-8B, which matches the performance of DeepSeek-R1 on math tasks. This model has 760M active parameters. The release of ZAYA1-8B is significant for the AI community as it provides a new, highly performant model for math-related tasks. Engineers can explore and utilize ZAYA1-8B for their projects. The model's open-source nature allows for further development and improvement.

Source →
FeedLens — Signal over noise Last 7 days