Quick start

ollama run gemma3:latest

Runnable variants

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

12 of 29 matching variants shown.29 of 29 variants shown.

Tag	Type	Inputs	Digest	Download	Parameters	Quantization	Context	Est. memory	Run
12b	Typelocal	Inputstext, image	Digestb92b218582c5	Download8.2 GB	Parameters12b	QuantizationUnknown	Context128K	Est. memory≈ 10.5 GB	`ollama run gemma3:12b`
12b-it-bf16	Typelocal	Inputstext, image	Digest36949f7228f6	Download24 GB	Parameters12b	QuantizationBF16	Context128K	Est. memory≈ 28 GB	`ollama run gemma3:12b-it-bf16`
12b-it-fp16	Typelocal	Inputstext, image	Digest36949f7228f6	Download24 GB	Parameters12b	QuantizationFP16	Context128K	Est. memory≈ 28 GB	`ollama run gemma3:12b-it-fp16`
12b-it-q4_K_M	Typelocal	Inputstext, image	Digestb92b218582c5	Download8.2 GB	Parameters12b	QuantizationQ4_K_M	Context128K	Est. memory≈ 10.5 GB	`ollama run gemma3:12b-it-q4_K_M`
12b-it-q8_0	Typelocal	Inputstext, image	Digest88c2e99b4a44	Download13 GB	Parameters12b	QuantizationQ8_0	Context128K	Est. memory≈ 16 GB	`ollama run gemma3:12b-it-q8_0`
12b-it-qat	Typelocal	Inputstext, image	Digestc615ec41b385	Download8.9 GB	Parameters12b	QuantizationUnknown	Context128K	Est. memory≈ 11 GB	`ollama run gemma3:12b-it-qat`
1b	Typelocal	Inputstext	Digest97558784bdc9	Download806 MB	Parameters1b	QuantizationUnknown	Context32K	Est. memory≈ 2 GB	`ollama run gemma3:1b`
1b-it-fp16	Typelocal	Inputstext	Digestd685bd623bd3	Download2.0 GB	Parameters1b	QuantizationFP16	Context32K	Est. memory≈ 3.5 GB	`ollama run gemma3:1b-it-fp16`
1b-it-q4_K_M	Typelocal	Inputstext	Digest97558784bdc9	Download806 MB	Parameters1b	QuantizationQ4_K_M	Context32K	Est. memory≈ 2 GB	`ollama run gemma3:1b-it-q4_K_M`
1b-it-q8_0	Typelocal	Inputstext	Digeste3acf9b8e078	Download1.1 GB	Parameters1b	QuantizationQ8_0	Context32K	Est. memory≈ 2.5 GB	`ollama run gemma3:1b-it-q8_0`
1b-it-qat	Typelocal	Inputstext	Digestb491bd3989c6	Download1.0 GB	Parameters1b	QuantizationUnknown	Context32K	Est. memory≈ 2.5 GB	`ollama run gemma3:1b-it-qat`
270m	Typelocal	Inputstext	Digeste7d36fb2c3b3	Download292 MB	Parameters270m	QuantizationUnknown	Context32K	Est. memory≈ 1.5 GB	`ollama run gemma3:270m`
270m-it-bf16	Typelocal	Inputstext	Digestdc598b095ea6	Download543 MB	Parameters270m	QuantizationBF16	Context32K	Est. memory≈ 2 GB	`ollama run gemma3:270m-it-bf16`
270m-it-fp16	Typelocal	Inputstext	Digesta44af03dd6b3	Download543 MB	Parameters270m	QuantizationFP16	Context32K	Est. memory≈ 2 GB	`ollama run gemma3:270m-it-fp16`
270m-it-q8_0	Typelocal	Inputstext	Digeste7d36fb2c3b3	Download292 MB	Parameters270m	QuantizationQ8_0	Context32K	Est. memory≈ 1.5 GB	`ollama run gemma3:270m-it-q8_0`
270m-it-qat	Typelocal	Inputstext	Digestb16d6d39dfbd	Download241 MB	Parameters270m	QuantizationUnknown	Context32K	Est. memory≈ 1.5 GB	`ollama run gemma3:270m-it-qat`
27b	Typelocal	Inputstext, image	Digest6be2e2ba6843	Download17 GB	Parameters27b	QuantizationUnknown	Context128K	Est. memory≈ 20 GB	`ollama run gemma3:27b`
27b-it-bf16	Typelocal	Inputstext, image	Digest1bc907263cde	Download55 GB	Parameters27b	QuantizationBF16	Context128K	Est. memory≈ 62.5 GB	`ollama run gemma3:27b-it-bf16`
27b-it-fp16	Typelocal	Inputstext, image	Digest1bc907263cde	Download55 GB	Parameters27b	QuantizationFP16	Context128K	Est. memory≈ 62.5 GB	`ollama run gemma3:27b-it-fp16`
27b-it-q4_K_M	Typelocal	Inputstext, image	Digest6be2e2ba6843	Download17 GB	Parameters27b	QuantizationQ4_K_M	Context128K	Est. memory≈ 20 GB	`ollama run gemma3:27b-it-q4_K_M`
27b-it-q8_0	Typelocal	Inputstext, image	Digest62dc427fec6b	Download30 GB	Parameters27b	QuantizationQ8_0	Context128K	Est. memory≈ 35 GB	`ollama run gemma3:27b-it-q8_0`
27b-it-qat	Typelocal	Inputstext, image	Digesta53d94b409fe	Download18 GB	Parameters27b	QuantizationUnknown	Context128K	Est. memory≈ 21.5 GB	`ollama run gemma3:27b-it-qat`
4b	Typelocal	Inputstext, image	Digest68ceab7e468a	Download3.4 GB	Parameters4b	QuantizationUnknown	Context128K	Est. memory≈ 5 GB	`ollama run gemma3:4b`
4b-it-bf16	Typelocal	Inputstext, image	Digest4e0b0e7a1c0a	Download8.6 GB	Parameters4b	QuantizationBF16	Context128K	Est. memory≈ 11 GB	`ollama run gemma3:4b-it-bf16`
4b-it-fp16	Typelocal	Inputstext, image	Digest4e0b0e7a1c0a	Download8.6 GB	Parameters4b	QuantizationFP16	Context128K	Est. memory≈ 11 GB	`ollama run gemma3:4b-it-fp16`
4b-it-q4_K_M	Typelocal	Inputstext, image	Digest68ceab7e468a	Download3.4 GB	Parameters4b	QuantizationQ4_K_M	Context128K	Est. memory≈ 5 GB	`ollama run gemma3:4b-it-q4_K_M`
4b-it-q8_0	Typelocal	Inputstext, image	Digest69f0d5459ab0	Download5.0 GB	Parameters4b	QuantizationQ8_0	Context128K	Est. memory≈ 7 GB	`ollama run gemma3:4b-it-q8_0`
4b-it-qat	Typelocal	Inputstext, image	Digestb0313423c944	Download4.0 GB	Parameters4b	QuantizationUnknown	Context128K	Est. memory≈ 5.5 GB	`ollama run gemma3:4b-it-qat`
latestdefault	Typelocal	Inputstext, image	Digest68ceab7e468a	Download3.4 GB	ParametersUnknown	QuantizationUnknown	Context128K	Est. memory≈ 5 GB	`ollama run gemma3:latest`

Related models

embeddinggemma-2local

EmbeddingGemma 2 is a multimodal embedding model from Google built on the Gemma 4 architecture.

bonsai-2local

Bonsai 2 from Prism ML is a family of vision models built for ternary weights – smaller models with close to full-precision quality.

qwen3.5local

Qwen 3.5 is a family of open-source multimodal models that delivers exceptional utility and performance.

moondreamlocal

moondream2 is a small vision language model designed to run efficiently on edge devices.