deepcogito/cogito-v1-preview-llama-3B
TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kPublished:Mar 31, 2025License:llama3.2Architecture:Transformer0.1K Warm

The deepcogito/cogito-v1-preview-llama-3B is a 3.2 billion parameter instruction-tuned generative language model developed by DeepCogito. It is a hybrid reasoning model capable of direct answers or self-reflection, optimized for coding, STEM, instruction following, and general helpfulness. Trained using Iterated Distillation and Amplification (IDA) across over 30 languages, it supports a 128k token context length and excels in multilingual, coding, and tool-calling capabilities compared to size-equivalent counterparts.

Loading preview...