DeepSeek V3
671.1B total, 37.6B active · 160k context · DeepSeek Model License (Model Agreement) · released 2024-12-25
chatcodingreasoning
What to run it on in Canada
Fully in memory at 64k 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
Only FP8 weights are provided (BF16 requires a conversion script); the HF repo totals 685B including 14B of MTP module weights. Hugging Face Transformers is not directly supported; use SGLang, vLLM, LMDeploy or TensorRT-LLM. Excludes 13.4B parameters of multi-token-prediction layers, which local runtimes do not load.
Questions
- What licence is DeepSeek V3 released under?
- DeepSeek Model License (Model Agreement), which allows commercial use.