DeepSeek-V3.2, a model that harmonizes high computational efficiency with superior reasoning and agent performance
DeepSeek-V3.2 Speciale, a model that harmonizes high computational efficiency with superior reasoning and agent performance
DeepSeek-V3.1 is a hybrid model that enhances tool usage, thinking efficiency, and supports both thinking and non-thinking modes via chat template switching
DeepSeek-V3-0324 demonstrates notable improvements over its predecessor, DeepSeek-V3, in several key aspects, including enhanced reasoning, improved function calling, and superior code generation capabilities.
The DeepSeek R1 0528 model has improved reasoning capabilities, this version also offers a reduced hallucination rate, enhanced support for function calling, and better experience for vibe coding.
A strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.
DeepSeek-R1 excels at reasoning tasks using a step-by-step training process, such as language, scientific reasoning, and coding tasks.