Skip to content
localmodel.run

Memory budget · 8 GB

Best local LLMs for 8GB

8GB is not a single ceiling. An 8GB Mac and an 8GB Laptop each leave a different amount free for model weights, so the largest model you can run changes with the memory type, not just the number.

Usable range
5–5.5 GB
Models that fit
24
Memory types
2
Top pick
4B

What 8GB actually gives you

Apple unified memory ~5.5 GB usable on Apple M1 (8GB) Comfortable ceiling: 4B System RAM (CPU only) ~5 GB usable on 8GB RAM Laptop (CPU/iGPU only) Comfortable ceiling: 4B

Usable figures are sourced per device (tap a card for the full profile). Verdicts below use Q4_K_M, the community-default quant.

Top pick for 8GB Q4_K_M
Gemma 3 4B 4B

Runs comfortably on the most capable 8GB setup (Apple M1 (8GB), ~5.5 GB usable) at ~3.8 GB. Check it against your exact device on its model page.

Models ranked for 8GB

Biggest that fits first Mac · Laptop

Each chip links to the full breakdown for that model on a real 8GB device. "Tight" means it fits but with little headroom, close other apps.

The ceiling, per memory type

Apple M1 (8GB) (~5.5 GB usable)

Runs up to Qwen3 4B (4B) comfortably at Q4_K_M. Larger models either sit tight or spill past the ~5.5 GB it can give a model.

8GB RAM Laptop (CPU/iGPU only) (~5 GB usable)

Runs up to Qwen3 4B (4B) comfortably at Q4_K_M. Larger models either sit tight or spill past the ~5 GB it can give a model.

8GB phones & tablets

Phones report 8GB too, but iOS/Android reserve more and the runtimes differ. Their usable pool is smaller:

iPhone 15 Pro ~4.5 GB iPhone 16 ~4.5 GB iPhone 16 Pro ~4.5 GB Generic Android Phone (8GB RAM) ~4.5 GB iPhone 17 ~4.5 GB

Too large for any 8GB device

Mistral 7B 7B Qwen2.5 7B 7B DeepSeek-R1-Distill-Qwen 7B 7B Qwen2.5 Coder 7B 7B Llama 3.1 8B 8B Qwen3 8B 8B DeepSeek-R1-Distill-Llama 8B 8B Qwen2.5-VL 7B 8.29B Gemma 2 9B 9B GLM-4 9B 9B Falcon3 10B 10B Llama 3.2 Vision 11B 10.7B Gemma 3 12B 12B Mistral Nemo 12B 12.2B Phi-4 14B 14B Qwen2.5 14B 14B Qwen3 14B 14B DeepSeek-R1-Distill-Qwen 14B 14B Qwen2.5 Coder 14B 14B DeepSeek-V2-Lite 16B gpt-oss 20B 21B Mistral Small 3 24B 24B Sarvam-M 24B 24B Gemma 2 27B 27B Gemma 3 27B 27B Sarvam-30B 30B Qwen3 30B-A3B 30.5B Qwen2.5 32B 32B Qwen3 32B 32B DeepSeek-R1-Distill-Qwen 32B 32B Qwen2.5 Coder 32B 32B Granite 4.0 H Small 32B Yi 1.5 34B 34B Command R 35B 35B Mixtral 8x7B 46.7B Llama 3.3 70B 70B Qwen2.5 72B 72B Sarvam-105B 105B Llama 4 Scout 109B gpt-oss 120B 117B Qwen3 235B A22B 235B DeepSeek R1 671B DeepSeek V3 671B

FAQ

How much of 8GB can a model actually use?

It depends on the memory type. Apple unified memory: about 5.5 GB (Apple M1 (8GB)); System RAM (CPU only): about 5 GB (8GB RAM Laptop (CPU/iGPU only)). The rest is reserved for the OS, display and runtime overhead.

What is the best local LLM for 8GB?

Gemma 3 4B (4B) is the strongest model that runs comfortably at Q4_K_M on the most capable 8GB setup (Apple M1 (8GB), ~5.5 GB usable). On a tighter 8GB device the ceiling is lower, shown per row above.

Sources

Memory figures are estimates at Q4_K_M with a small context. See methodology.