deepcogito/cogito-v2-preview-llama-70B

The deepcogito/cogito-v2-preview-llama-70B is a 70 billion parameter instruction-tuned generative language model developed by DeepCogito. This model is a hybrid reasoning model, capable of both direct answers and self-reflection, trained using Iterated Distillation and Amplification (IDA) for iterative self-improvement. It is optimized for coding, STEM, instruction following, and general helpfulness, offering significantly enhanced multilingual, coding, and tool-calling capabilities compared to similarly sized counterparts. The model supports a context length of 128k tokens and is trained in over 30 languages.