According to Deepseek.com news, DeeepSeek released a preview version of DeepSeek-V4 series on April 4. It includs two strong Mixture-of-Experts (MoE) language models, DeepSeek-V4-Pro with 1.6T parameters and DeepSeek-V4-Flash with 284B parameters. Both versions support a context length of one million tokens.
For further detail information, you could get the official Chinese document from DeepSeek-V4 预览版：迈入百万上下文普惠时代 (qq.com). Or you could get the English document from deepseek-ai/DeepSeek-V4-Pro · Hugging Face.