Expired50% confidencePredictionExact time
Subsequent versions of DeepSeek V4 will support native multimodality, but multimodality is only one component; the main focus remains on improving AI training and solving continual learning.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/25/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/23/2026(expired)
Sources
梁文锋深度自白：DeepSeek的克制哲学、开源策略与AGI路线图
Related Claims
UnverifiedDeepSeek V4采用MUON优化器替代AdamW进行训练77% similarUnverifiedThe core innovations of DeepSeek V4 include continued optimization of the MoE (Mixture of Experts) architecture and deep engineering tuning of the training pipeline.74% similarUnverifiedThe training cost of the DeepSeek V3 series is only a fraction of that of comparable models.74% similarUnverifiedDeepSeek V4 performed steadily in the full-stack coding test with largely complete functionality, but had shortcomings in some animations, ranking second.73% similarUnverifiedDeepSeek will not make video generation or world models its main direction, but versions after V4 will support native multimodality.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/615533API
curl https://kongchang.com/api/v1/knowledge/claims/615533MCP
get_claim(id=615533)