Back to Calculator Deploy on RunPodDeploy Now
DeepSeek V4
Hot~1 trillion total parameter MoE model from DeepSeek with ~32B active per token. Introduces Engram conditional memory (O(1) knowledge retrieval), mHC training stability, and sparse attention. 1M token context, native multimodal (text, image, video). Optimized for Huawei Ascend chips. Apache 2.0 expected.
Specifications
SourceArchitectureVISION
Parameters1000B
Familydeepseek
VRAM (Q4)500.0G
MoE: 32B active.
Expected late April 2026. V4 Lite accessible on API since early April 2026.
deepseekmoereasoningcodingmultimodaltrendingpreviewapache2
Run in the Cloud
This model requires enterprise-grade VRAM. Rent GPUs on RunPod and start generating.
Instant Cloud GPUs
Running out of VRAM? Rent a high-end H100 or RTX 4090 on RunPod and deploy in seconds.
Share this Model
Send these specs directly to your community.