Quick start

ollama run d3:2b-mxfp8

Runnable variants

Memory is a rough file-size estimate; context length, runtime and GPU offload change it.

12 of 16 matching variants shown.16 of 16 variants shown.

Tag	Type	Inputs	Digest	Download	Parameters	Quantization	Context	Est. memory	Run
27b	Typelocal	Inputstext, image	Digestbc873824c9ed	DownloadUnknown	Parameters27b	QuantizationUnknown	Context256K	Est. memoryUnknown	`ollama run d3:27b`
27b-mxfp8	Typelocal	Inputstext, image	Digestd3388a380a7c	Download29 GB	Parameters27b	QuantizationMXFP8	Context256K	Est. memory≈ 33.5 GB	`ollama run d3:27b-mxfp8`
27b-nvfp4	Typelocal	Inputstext, image	Digestb93588ba7385	Download17 GB	Parameters27b	QuantizationNVFP4	Context256K	Est. memory≈ 20 GB	`ollama run d3:27b-nvfp4`
27b-q8_0	Typelocal	Inputstext, image	Digestbdde0a6212d6	Download30 GB	Parameters27b	QuantizationQ8_0	Context256K	Est. memory≈ 35 GB	`ollama run d3:27b-q8_0`
2b	Typelocal	Inputstext, image	Digestb31336d5e0cc	DownloadUnknown	Parameters2b	QuantizationUnknown	Context256K	Est. memoryUnknown	`ollama run d3:2b`
2b-mxfp8	Typelocal	Inputstext, image	Digestdbf58974e0fe	Download3.1 GB	Parameters2b	QuantizationMXFP8	Context256K	Est. memory≈ 4.5 GB	`ollama run d3:2b-mxfp8`
2b-q8_0	Typelocal	Inputstext, image	Digeste1b3991aa480	Download3.2 GB	Parameters2b	QuantizationQ8_0	Context256K	Est. memory≈ 5 GB	`ollama run d3:2b-q8_0`
4b	Typelocal	Inputstext, image	Digest50e469b6b0e6	DownloadUnknown	Parameters4b	QuantizationUnknown	Context256K	Est. memoryUnknown	`ollama run d3:4b`
4b-mxfp8	Typelocal	Inputstext, image	Digest9cd750f6fa24	Download5.6 GB	Parameters4b	QuantizationMXFP8	Context256K	Est. memory≈ 7.5 GB	`ollama run d3:4b-mxfp8`
4b-nvfp4	Typelocal	Inputstext, image	Digestde276ebc958a	Download4.0 GB	Parameters4b	QuantizationNVFP4	Context256K	Est. memory≈ 5.5 GB	`ollama run d3:4b-nvfp4`
4b-q8_0	Typelocal	Inputstext, image	Digest2c571f91caf6	Download5.8 GB	Parameters4b	QuantizationQ8_0	Context256K	Est. memory≈ 7.5 GB	`ollama run d3:4b-q8_0`
9b	Typelocal	Inputstext, image	Digest6e6c5ab6ce3a	Download10 GB	Parameters9b	QuantizationUnknown	Context256K	Est. memory≈ 12.5 GB	`ollama run d3:9b`
9b-mxfp8	Typelocal	Inputstext, image	Digest518acc51965d	Download10 GB	Parameters9b	QuantizationMXFP8	Context256K	Est. memory≈ 12.5 GB	`ollama run d3:9b-mxfp8`
9b-nvfp4	Typelocal	Inputstext, image	Digest9816d390dd1f	Download7.7 GB	Parameters9b	QuantizationNVFP4	Context256K	Est. memory≈ 10 GB	`ollama run d3:9b-nvfp4`
9b-q8_0	Typelocal	Inputstext, image	Digest0f459473f035	Download10 GB	Parameters9b	QuantizationQ8_0	Context256K	Est. memory≈ 12.5 GB	`ollama run d3:9b-q8_0`
latest	Typelocal	Inputstext, image	Digest50e469b6b0e6	DownloadUnknown	ParametersUnknown	QuantizationUnknown	Context256K	Est. memoryUnknown	`ollama run d3:latest`

Related models

llavalocal

🌋 LLaVA is a novel end-to-end trained large multimodal model that combines a vision encoder and Vicuna for general-purpose visual and language understanding. Updated to version 1.6.

qwen2.5vllocal

Flagship vision-language model of Qwen and also a significant leap from the previous Qwen2-VL.

ministral-3local

The Ministral 3 family is designed for edge deployment, capable of running on a wide range of hardware.

translategemmalocal

A new collection of open translation models built on Gemma 3, helping people communicate across 55 languages.