The latest members of the Olmo 3 family: another 3 weeks of RL for 32B Think, the 32B Instruct model, large post-training research datasets...
-
allenai/Olmo-3.1-32B-Think
Text Generation β’ 32B β’ Updated β’ 4.35k β’ β’ 64 -
allenai/Olmo-3.1-32B-Instruct-SFT
32B β’ Updated β’ 2.31k β’ 6 -
allenai/Olmo-3.1-32B-Instruct-DPO
Text Generation β’ 32B β’ Updated β’ 1.02k β’ 4 -
allenai/Olmo-3.1-32B-Instruct
Text Generation β’ 32B β’ Updated β’ 7.58k β’ β’ 44