1,598  Downloads Updated  1 year ago

32B reasoning model trained from Qwen2.5-32B-Instruct with 17K data with performance on par with o1 preview.