Multimodal LLMs
21models · local AI VRAM requirements & GPU compatibility
Multimodal models understand images, audio, or video in addition to text. Running them locally requires the same VRAM as their text-only counterparts for the language backbone, plus additional memory for the vision encoder during inference. Check each model page for the exact VRAM breakdown.
- Kimi K3Moonshot AI · 2800B params (104B active)1910.3 GBQ4_K_M
- DeepSeek V4 Pro 1.6TDeepSeek · 1600B params (49B active)1092.5 GBQ4_K_M
- Kimi K2.6Moonshot AI · 1000B params (32B active)685.1 GBQ4_K_M
- Kimi K2.5Moonshot AI · 1000B params (32B active)683.2 GBQ4_K_M
- InklingThinking Machines Lab · 975B params (41B active)667.5 GBQ4_K_M
- MiniMax M3MiniMax · 428B params (23B active)293.1 GBQ4_K_M
- Llama 4 Maverick 400BMeta · 400B params (17B active)277.3 GBQ4_K_M
- DeepSeek V4 Flash 284BDeepSeek · 284B params (13B active)194.5 GBQ4_K_M
- Step 3.7 FlashStepFun · 198B params (11B active)136.7 GBQ4_K_M
- Llama 4 Scout 109BMeta · 109B params (17B active)77.4 GBQ4_K_Mfits 80 GB
- GLM-4.6V 106BZ.ai · 106B params (12B active)74.0 GBQ4_K_Mfits 80 GB
- Gemma 4 31BGoogle · 31B params24.8 GBQ4_K_Mfits 48 GB
- Gemma 3 27B InstructGoogle · 27B params20.1 GBQ4_K_Mfits 24 GB
- Bonsai 27BPrismML · 27B params6.1 GBQ4_K_Mfits 8 GB
- Gemma 4 26B (MoE)Google · 26B params (3.8B active)19.3 GBQ4_K_Mfits 24 GB
- Mistral Small 3.1 24B InstructMistral AI · 24B params17.9 GBQ4_K_Mfits 24 GB
- Gemma 3 12B InstructGoogle · 12.2B params9.5 GBQ4_K_Mfits 12 GB
- Gemma 4 12B (Unified)Google · 12B params11.8 GBQ4_K_Mfits 12 GB
- Gemma 3 4B InstructGoogle · 4B params3.3 GBQ4_K_Mfits 8 GB
- Gemma 4 E4BGoogle · 4B params3.9 GBQ4_K_Mfits 8 GB
- Gemma 4 E2BGoogle · 2B params1.8 GBQ4_K_Mfits 8 GB
Want to check your specific GPU? Use the homepage calculator to see which of these models fit your hardware with estimated tokens per second.