DeepSeek V4 Pro 0813
1623B total, 99.8B active · 1024k context · MIT License · released 2026-08-13
chatcodingagenticreasoning
What to run it on in Germany
Fully in memory at 32k context, priced at retailers today.
Nothing we price holds this model fully in memory.
All hardware that runs it
Everything we price, best value first. Two cards split the model by layers.
Nothing we price can run it at this context.
Memory needed
Weights, KV cache and runtime buffers at each context length (KV cache at fp16).
| Quant | File |
|---|
Run it
Copy a command for the runtime you use.
No published builds to run yet.
Notes
No Jinja chat template; use the provided Python `encoding` folder. Supports DSpark speculative decoding in vLLM and SGLang; the vLLM example runs on a 4×GB300 node. Sparse or compressed attention: KV read per token is estimated as full attention. deepseek_v4 layout is new; verify against measured results. Excludes 27.5B parameters of multi-token-prediction layers, which local runtimes do not load.
Questions
- What licence is DeepSeek V4 Pro 0813 released under?
- MIT License, which allows commercial use.