Quick start

ollama run granite3-moe:latest

Runnable variants

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

12 of 33 matching variants shown.33 of 33 variants shown.

Tag	Type	Inputs	Digest	Download	Parameters	Quantization	Context	Est. memory	Run
1b	Typelocal	Inputstext	Digestd84e1e38ee39	Download822 MB	Parameters1b	QuantizationUnknown	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b`
1b-instruct-fp16	Typelocal	Inputstext	Digestcee5d696574c	Download2.7 GB	Parameters1b	QuantizationFP16	Context4K	Est. memory≈ 4.5 GB	`ollama run granite3-moe:1b-instruct-fp16`
1b-instruct-q2_K	Typelocal	Inputstext	Digest5965f707577c	Download512 MB	Parameters1b	QuantizationQ2_K	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q2_K`
1b-instruct-q3_K_L	Typelocal	Inputstext	Digestbbb955565d3d	Download711 MB	Parameters1b	QuantizationQ3_K_L	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q3_K_L`
1b-instruct-q3_K_M	Typelocal	Inputstext	Digestea6a79b9b3cc	Download659 MB	Parameters1b	QuantizationQ3_K_M	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q3_K_M`
1b-instruct-q3_K_S	Typelocal	Inputstext	Digestcab9e1d32d19	Download598 MB	Parameters1b	QuantizationQ3_K_S	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q3_K_S`
1b-instruct-q4_0	Typelocal	Inputstext	Digest9b6ee4c08df8	Download768 MB	Parameters1b	QuantizationQ4_0	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q4_0`
1b-instruct-q4_1	Typelocal	Inputstext	Digestc3aa5d722767	Download849 MB	Parameters1b	QuantizationQ4_1	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q4_1`
1b-instruct-q4_K_M	Typelocal	Inputstext	Digestd84e1e38ee39	Download822 MB	Parameters1b	QuantizationQ4_K_M	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q4_K_M`
1b-instruct-q4_K_S	Typelocal	Inputstext	Digest76e6392d08a3	Download775 MB	Parameters1b	QuantizationQ4_K_S	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:1b-instruct-q4_K_S`
1b-instruct-q5_0	Typelocal	Inputstext	Digestf49b9c68d958	Download929 MB	Parameters1b	QuantizationQ5_0	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:1b-instruct-q5_0`
1b-instruct-q5_1	Typelocal	Inputstext	Digestf1dbdb4ce40f	Download1.0 GB	Parameters1b	QuantizationQ5_1	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:1b-instruct-q5_1`
1b-instruct-q5_K_M	Typelocal	Inputstext	Digestc51ea1c8bab3	Download956 MB	Parameters1b	QuantizationQ5_K_M	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:1b-instruct-q5_K_M`
1b-instruct-q5_K_S	Typelocal	Inputstext	Digest427124d0793f	Download929 MB	Parameters1b	QuantizationQ5_K_S	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:1b-instruct-q5_K_S`
1b-instruct-q6_K	Typelocal	Inputstext	Digesta585cc84307b	Download1.1 GB	Parameters1b	QuantizationQ6_K	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:1b-instruct-q6_K`
1b-instruct-q8_0	Typelocal	Inputstext	Digest84f7efa9cc2d	Download1.4 GB	Parameters1b	QuantizationQ8_0	Context4K	Est. memory≈ 3 GB	`ollama run granite3-moe:1b-instruct-q8_0`
3b	Typelocal	Inputstext	Digest157f538ae66e	Download2.1 GB	Parameters3b	QuantizationUnknown	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b`
3b-instruct-fp16	Typelocal	Inputstext	Digest9433d98bf426	Download6.8 GB	Parameters3b	QuantizationFP16	Context4K	Est. memory≈ 9 GB	`ollama run granite3-moe:3b-instruct-fp16`
3b-instruct-q2_K	Typelocal	Inputstext	Digest48ee4c1b7090	Download1.3 GB	Parameters3b	QuantizationQ2_K	Context4K	Est. memory≈ 2.5 GB	`ollama run granite3-moe:3b-instruct-q2_K`
3b-instruct-q3_K_L	Typelocal	Inputstext	Digest8c76215f504a	Download1.8 GB	Parameters3b	QuantizationQ3_K_L	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b-instruct-q3_K_L`
3b-instruct-q3_K_M	Typelocal	Inputstext	Digestcc32cde7c1fe	Download1.6 GB	Parameters3b	QuantizationQ3_K_M	Context4K	Est. memory≈ 3 GB	`ollama run granite3-moe:3b-instruct-q3_K_M`
3b-instruct-q3_K_S	Typelocal	Inputstext	Digest5952b23cdd5c	Download1.5 GB	Parameters3b	QuantizationQ3_K_S	Context4K	Est. memory≈ 3 GB	`ollama run granite3-moe:3b-instruct-q3_K_S`
3b-instruct-q4_0	Typelocal	Inputstext	Digestd144aac2b0d8	Download1.9 GB	Parameters3b	QuantizationQ4_0	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b-instruct-q4_0`
3b-instruct-q4_1	Typelocal	Inputstext	Digestac91c7b62cc7	Download2.1 GB	Parameters3b	QuantizationQ4_1	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b-instruct-q4_1`
3b-instruct-q4_K_M	Typelocal	Inputstext	Digest157f538ae66e	Download2.1 GB	Parameters3b	QuantizationQ4_K_M	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b-instruct-q4_K_M`
3b-instruct-q4_K_S	Typelocal	Inputstext	Digest9f5ecea51621	Download1.9 GB	Parameters3b	QuantizationQ4_K_S	Context4K	Est. memory≈ 3.5 GB	`ollama run granite3-moe:3b-instruct-q4_K_S`
3b-instruct-q5_0	Typelocal	Inputstext	Digeste623edd5db90	Download2.3 GB	Parameters3b	QuantizationQ5_0	Context4K	Est. memory≈ 4 GB	`ollama run granite3-moe:3b-instruct-q5_0`
3b-instruct-q5_1	Typelocal	Inputstext	Digest709d47478efd	Download2.5 GB	Parameters3b	QuantizationQ5_1	Context4K	Est. memory≈ 4 GB	`ollama run granite3-moe:3b-instruct-q5_1`
3b-instruct-q5_K_M	Typelocal	Inputstext	Digest94e89b071306	Download2.4 GB	Parameters3b	QuantizationQ5_K_M	Context4K	Est. memory≈ 4 GB	`ollama run granite3-moe:3b-instruct-q5_K_M`
3b-instruct-q5_K_S	Typelocal	Inputstext	Digestc84b2a3a27fd	Download2.3 GB	Parameters3b	QuantizationQ5_K_S	Context4K	Est. memory≈ 4 GB	`ollama run granite3-moe:3b-instruct-q5_K_S`
3b-instruct-q6_K	Typelocal	Inputstext	Digestc196db4b29bb	Download2.8 GB	Parameters3b	QuantizationQ6_K	Context4K	Est. memory≈ 4.5 GB	`ollama run granite3-moe:3b-instruct-q6_K`
3b-instruct-q8_0	Typelocal	Inputstext	Digestf3b962d19110	Download3.6 GB	Parameters3b	QuantizationQ8_0	Context4K	Est. memory≈ 5.5 GB	`ollama run granite3-moe:3b-instruct-q8_0`
latestdefault	Typelocal	Inputstext	Digestd84e1e38ee39	Download822 MB	ParametersUnknown	QuantizationUnknown	Context4K	Est. memory≈ 2 GB	`ollama run granite3-moe:latest`

Source-listed applications

Claude CodeHermes AgentOpenClawOpenCode

Related models

qwen3local

Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.

granite3.1-moelocal

The IBM Granite 1B and 3B models are long-context mixture of experts (MoE) Granite models from IBM designed for low latency usage.

llama3.2local

Meta's Llama 3.2 goes small with 1B and 3B models.

granite4local

Granite 4 features improved instruction following (IF) and tool-calling capabilities, making them more effective in enterprise applications.