DeepSeek-VL2
MoE vision-language, now easier to access. DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding. DeepSeek-VL2 are open-source vision-language models with strong multimodal understanding, powered by an efficient MoE architecture.
Open source
Reactions (1)
Similar products
by description similarity
DeepSeek-V4
huggingface.coThe open-source era of 1M context intelligence
DeepSeek R1
github.comAdvanced reasoning model
DeepSeek-R1-0528
huggingface.coNew open-source LLM that rivals o3 in coding & reasoning
DeepSeek-V3.1
chat.deepseek.comOur first step toward the agent era
Wan 2.2
wan.videoThe first open MoE model for AI video generation