Cost to run
DeepSeek V3-0324
deepseek/deepseek-v3-0324
- Family
- DeepSeek V3
- Context
- 131,072 tokens
DeepSeek V3-0324 can be run self-hosted (rent a GPU + run vLLM/TGI) or through a serverless API (pay per token). Live pricing comparisons:
Self-hosted
Rent a GPU from Lambda, RunPod, Vast.ai, CoreWeave. Full control.
Serverless
Pay per token via Together, Fireworks, DeepInfra, Groq.
Need to benchmark this model against another? Try the calculator or see where it ranks on the InferenceScore leaderboard.