Olmo Hybrid

Olmo Hybrid

7B open model mixing transformers and linear RNNs

Product HuntOpen SourceLaunched Mar 7, 2026107#11 of the day

About Olmo Hybrid

Olmo Hybrid is a fully open 7B model that combines transformer attention with linear RNN layers. Utilizing a 3:1 pattern of Gated DeltaNet to attention, it matches the accuracy of Olmo 3 on MMLU while using 49% fewer tokens.

The person behind Olmo Hybrid

Olmo Hybrid's team page →

More Open Source launches

Made Olmo Hybrid? Ask us to remove this listing or your details.