RTX 4090 Desktop (24 GB VRAM, 64 GB RAM) — Local LLM Performance

RTX 4090 Desktop (24 GB VRAM, 64 GB RAM) is a Desktop running Windows or Linux, with 24 GB VRAM + 64 GB system RAM.

Specifications

Memory for models	24 GB
Memory bandwidth	1,008 GB/s
Memory upgradeable	Yes
Chip	NVIDIA GeForce RTX 4090
Architecture	Ada Lovelace AD102
Form factor	Desktop
Operating system	Windows or Linux
Released	2022
Llama 3.1 8B Q4_K_M	86–165 tok/s

Worth knowing before you buy

• 24 GB is the sweet spot for 27–32B models at Q4; 70B needs offload or a second card.

Models on the RTX 4090 desktop

• Yi 1.5 Family — Yes, but it is tight
• VibeThinker — Yes — comfortably
• Aya 3B (Tiny Aya) — Yes — comfortably
• StarCoder 2 — Yes — comfortably
• SmolLM3 — Yes — comfortably
• SmolLM2 — Yes — comfortably
• Qwen3-Coder — Yes — comfortably
• Qwen3.8 — Yes, but it is tight
• Qwen 3.7 — Yes, but it is tight
• Qwen 3.6 — Yes, but it is tight
• Qwen 3.5 — Yes, but it is tight
• Qwen 3 — Yes, but it is tight
• Qwen 2.5 VL — Yes — comfortably
• Qwen 2.5 Family — Yes, but it is tight
• Poolside Laguna XS 2.1 — Yes, but it is tight
• Phi-4 Mini — Yes — comfortably
• Phi-4 Family — Yes — comfortably
• Phi 3.5 Family — Yes — comfortably
• OLMo 2 — Yes, but it is tight
• Nemotron Cascade 2 — Yes, but it is tight
• Nemotron 3 Nano Omni — Yes, but it is tight
• Mistral Small 3.1 — Yes
• Mistral Small 3.2 — Yes, but it is tight
• Mistral Family — Yes, but it is tight
• Ministral — Yes — comfortably
• Ministral 3 — Yes — comfortably
• Magistral Small — Yes, but it is tight
• Llama 3.2 Vision — Yes — comfortably
• Llama 3.2 Family — Yes — comfortably
• Llama 3.1 Family — Yes — comfortably
• InternLM 3 — Yes, but it is tight
• IBM Granite 4.0 — Yes, but it is tight
• IBM Granite 4.1 — Yes, but it is tight
• Granite 3.0 — Yes — comfortably
• GPT-OSS — Yes, but it is tight
• GLM-4.7 / GLM-Z1 — Yes, but it is tight
• GLM-6 — Yes — comfortably
• GLM-5 / GLM-5.1 — Yes, but it is tight
• Gemma 4 — Yes
• Gemma 3n — Yes — comfortably
• Gemma 3 — Yes — comfortably
• Gemma 2 Family — Yes — comfortably
• Falcon 3 — Yes — comfortably
• EXAONE 3.5 — Yes, but it is tight
• Devstral — Yes, but it is tight
• DeepSeek R1 — Yes, but it is tight
• Cosmos 3 — Yes
• Command R Family — Yes, but it is tight
• Cogito v1 — Yes, but it is tight
• Codestral — Yes
• Bonsai 27B — Yes, but it is tight
• BitNet b1.58 — Yes — comfortably
• Aya Expanse — Yes

← All hardware | AI stations and mini-PCs | Check your hardware