Quick start
ollama run deepseek-v3:latestRunnable variants
Memory is a rough file-size estimate; context length, runtime and GPU offload change it.
5 of 5 matching variants shown.5 of 5 variants shown.
| Tag | Type | Inputs | Digest | Download | Parameters | Quantization | Context | Est. memory | Run |
|---|---|---|---|---|---|---|---|---|---|
| 671b | Typelocal | Inputstext | Digest5da0e2d4a9e0 | Download404 GB | Parameters671b | QuantizationUnknown | Context160K | Est. memory≈ 453 GB | ollama run deepseek-v3:671b |
| 671b-fp16 | Typelocal | Inputstext | Digest7770bf5a5ed8 | Download1300 GB | Parameters671b | QuantizationFP16 | Context4K | Est. memory≈ 1454 GB | ollama run deepseek-v3:671b-fp16 |
| 671b-q4_K_M | Typelocal | Inputstext | Digest5da0e2d4a9e0 | Download404 GB | Parameters671b | QuantizationQ4_K_M | Context160K | Est. memory≈ 453 GB | ollama run deepseek-v3:671b-q4_K_M |
| 671b-q8_0 | Typelocal | Inputstext | Digest96061c74c1a5 | Download713 GB | Parameters671b | QuantizationQ8_0 | Context4K | Est. memory≈ 798 GB | ollama run deepseek-v3:671b-q8_0 |
| latestdefault | Typelocal | Inputstext | Digest5da0e2d4a9e0 | Download404 GB | ParametersUnknown | QuantizationUnknown | Context160K | Est. memory≈ 453 GB | ollama run deepseek-v3:latest |
Related models
deepseek-v3.1local
DeepSeek-V3.1-Terminus is a hybrid model that supports both thinking mode and non-thinking mode.
cogito-2.1local
The Cogito v2.1 LLMs are instruction tuned generative models. All models are released under MIT license for commercial use.
deepseek-v2.5local
An upgraded version of DeekSeek-V2 that integrates the general and coding abilities of both DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct.
qwen3.8-flash-nextlocal
This experimental preview of the architecture that will underpin Qwen4.