whatsmy.fyiYour digital footprint
← All modelsBROWSE BY RAM

Best local LLMs for 32 GB RAM

With 32 GB of total memory, 62 of the 73 models in our catalog run comfortably at the recommended 4-bit quantization. Top pick: Qwen 3.5 35B-A3B (21.2 GB download).

Reading your hardware signals…

Top picks

Best overallQwen 3.5 35B-A3BAlibaba Β· 35B Β· 21.2 GBCheck requirements β†’

All models that run on 32 GB

62 / 73 models
ModelParamsDownload (Q4)Min RAMBest for
Qwen 3.5 35B-A3B35B (A3B)21.2 GB32 GBChat, Reasoning, Coding, Vision
Qwen 3.6 35B-A3B35B (A3B)21.2 GB32 GBChat, Reasoning, Coding, Vision
Command R 35B35B21.2 GB32 GBChat
Qwen3-VL 32B33B20.0 GB32 GBVision, Chat, Reasoning
EXAONE 4.5 33B33B20.0 GB32 GBVision, Reasoning, Chat
Qwen 3 32B32.8B19.9 GB32 GBChat, Reasoning
Qwen 2.5 Coder 32B32.8B19.9 GB32 GBCoding
QwQ 32B32.8B19.9 GB32 GBReasoning
DeepSeek R1 32B32.8B19.9 GB32 GBReasoning
Granite 4.0 H Small32B (A9B)19.4 GB32 GBChat, Coding
Nemotron 3 Nano 30B-A3B31.6B (A3.6B)19.2 GB32 GBChat, Reasoning, Coding
Gemma 4 31B30.7B18.6 GB32 GBChat, Coding, Reasoning, Vision
Qwen 3 30B-A3B30.5B (A3.3B)18.5 GB32 GBChat, Reasoning
Qwen3-VL 30B-A3B30B (A3B)18.2 GB32 GBVision, Chat, Reasoning
Gemma 3 27B27.4B16.6 GB24 GBChat, Vision
Qwen 3.5 27B27B16.4 GB24 GBChat, Reasoning, Coding, Vision
Qwen 3.6 27B27B16.4 GB24 GBChat, Reasoning, Coding, Vision
Gemma 4 26B A4B25.2B (A3.8B)15.3 GB24 GBChat, Coding, Reasoning, Vision
Mistral Small 3.1 24B24B14.6 GB24 GBChat, Vision
Devstral 24B24B14.6 GB24 GBCoding
Magistral Small 1.224B14.6 GB24 GBReasoning, Chat, Vision
Devstral Small 2 24B24B14.6 GB24 GBCoding, Chat, Vision
Codestral 22B22.2B13.5 GB24 GBCoding
GPT-OSS 20B20.9B (A3.6B)12.7 GB24 GBChat, Reasoning
Phi-4 Reasoning Vision 15B15B9.1 GB16 GBVision, Reasoning
Qwen 3 14B14.8B9.0 GB16 GBChat, Reasoning
DeepSeek R1 14B14.8B9.0 GB16 GBReasoning
Phi-4 14B14.7B8.9 GB16 GBChat, Reasoning
Ministral 3 14B14B8.5 GB16 GBChat, Vision
OLMo 2 13B13.7B8.3 GB12 GBChat
Gemma 3 12B12.2B7.4 GB12 GBChat, Vision
Mistral Nemo 12B12.2B7.4 GB12 GBChat
Gemma 4 12B12B7.3 GB12 GBChat, Coding, Reasoning, Vision
Mellum 2 12B-A2.5B12B (A2.5B)7.3 GB12 GBCoding
Qwen 3.5 9B9B5.5 GB12 GBChat, Reasoning, Vision
GLM-4.6V-Flash9B5.5 GB12 GBVision, Chat
Qwen 2.5 VL 7B8.3B5.0 GB8 GBVision, Chat
Qwen 3 8B8.2B5.0 GB8 GBChat, Reasoning
Granite 3.3 8B8.2B5.0 GB8 GBChat
Llama 3.1 8B8B4.9 GB8 GBChat
DeepSeek R1 8B8B4.9 GB8 GBReasoning
Gemma 4 E4B8B (A4.5B)4.9 GB8 GBChat, Vision
Qwen3-VL 8B8B4.9 GB8 GBVision, Chat
Ministral 3 8B8B4.9 GB8 GBChat, Vision
Gemma 3n E4B7.8B (A4B)4.7 GB8 GBChat, Vision
Qwen 2.5 Coder 7B7.6B4.6 GB8 GBCoding
DeepSeek R1 7B7.6B4.6 GB8 GBReasoning
Mistral 7B7.2B4.4 GB8 GBChat
Gemma 4 E2B5.1B (A2.3B)3.1 GB6 GBChat, Vision
Gemma 3 4B4.3B2.6 GB6 GBChat, Vision
Qwen 3 4B4B2.4 GB6 GBChat, Reasoning
Qwen 3.5 4B4B2.4 GB6 GBChat, Vision
Phi-4 Mini 3.8B3.8B2.3 GB6 GBChat
Llama 3.2 3B3.2B1.9 GB4 GBChat
DeepSeek-OCR3B (A0.57B)1.8 GB4 GBVision
Ministral 3 3B3B1.8 GB4 GBChat, Vision
DeepSeek R1 1.5B1.8B1.1 GB3 GBReasoning
Qwen 3 1.7B1.7B1.0 GB3 GBChat
SmolLM2 1.7B1.7B1.0 GB3 GBChat
Llama 3.2 1B1.2B0.7 GB3 GBChat
Gemma 3 1B1B0.6 GB3 GBChat
Qwen 3 0.6B0.6B0.4 GB2 GBChat

Sizes are 4-bit (Q4_K_M) GGUF builds β€” the standard for running models locally. Β· Data updated: 2026-06-11 Β· How we calculate these numbers β†’

Just out of reach β€” needs 64 GB

These models don't fit comfortably in 32 GB but run on the next tier up.

Llama 3.1 70B Β· 64 GBLlama 3.3 70B Β· 64 GBDeepSeek R1 70B Β· 64 GBQwen3-Next 80B-A3B Β· 64 GBQwen3 Coder Next 80B-A3B Β· 64 GB
See the 64 GB tier β†’

Frequently asked questions

Browse by RAM

Best LLMs for 8 GB RAM

26 models run comfortably

Check requirements β†’

Best LLMs for 16 GB RAM

38 models run comfortably

Check requirements β†’

Best LLMs for 64 GB RAM

67 models run comfortably

Check requirements β†’
Best Local LLMs for 32 GB RAM (2026) β€” Tested Live in Your Browser