Best LLMs for 12 GB
Pick your memory. We rank every catalogued model that fits at its recommended quantization, by task, and show what the next tier up unlocks.
Written by Jakub Rusinowski · Last updated
Memory tier12 GBFor example Intel Arc B580, RTX 3060 (12GB), RX 7700 XT
Models that fit85entirely in memory
Best at 12 GB, by task
Score is the task-specific quality score on a 0–100 scale; speed is estimated for the Intel Arc B580.
CodingAll ranked →
1StarCoder 2 15BQ4_K_M · 9.4 GB942Qwen 3 14BQ4_K_M · 8.9 GB933Granite 3.0 8B InstructQ4_K_M · 4.8 GB92ReasoningAll ranked →
1Qwen 3 14BQ4_K_M · 8.9 GB942DeepSeek R1 Distill Qwen 14BQ4_K_M · 8.5 GB943Phi-4 (14B)Q4_K_M · 8.5 GB94The strongest that fit
Top 15, by logic and coding
| Model | Memory · marker = 12 GB | VRAM | Speed on Intel Arc B580 (est.) | Score |
|---|---|---|---|---|
| Qwen 2.5 14B Instruct Qwen 2.5 Family · Q4_K_M | 8.5 GB | ~26 tok/s | 95 | |
| Qwen 3 14B Qwen 3 · Q4_K_M | 8.9 GB | ~26 tok/s | 94 | |
| DeepSeek R1 Distill Qwen 14B DeepSeek R1 · Q4_K_M | 8.5 GB | ~27 tok/s | 92 | |
| Phi-4 (14B) Phi-4 Family · Q4_K_M | 8.5 GB | ~26 tok/s | 92 | |
| Granite 3.0 8B Instruct Granite 3.0 · Q4_K_M | 4.8 GB | ~41 tok/s | 90 | |
| Ternary Bonsai 27B Bonsai 27B · Ternary (1.58-bit, ~1.71 bpw) | 5.8 GB | ~35 tok/s | 90 | |
| Qwen 2.5 7B Instruct Qwen 2.5 Family · Q4_K_M | 4.6 GB | ~46 tok/s | 89 | |
| Qwen 3 8B Qwen 3 · Q4_K_M | 5 GB | ~41 tok/s | 89 | |
| DeepSeek R1 Distill Llama 8B DeepSeek R1 · Q4_K_M | 4.8 GB | ~42 tok/s | 89 | |
| Mistral NeMo 12B Mistral Family · Q4_K_M | 7.2 GB | ~30 tok/s | 88 | |
| Gemma 3 12B Instruct Gemma 3 · Q4_K_M | 7.2 GB | ~30 tok/s | 87 | |
| StarCoder 2 15B StarCoder 2 · Q4_K_M | 9.4 GB | ~25 tok/s | 86 | |
| InternLM 3 8B Instruct InternLM 3 · Q4_K_M | 5.3 GB | ~41 tok/s | 86 | |
| Qwen 3.5 9B Qwen 3.5 · Q4_K_M | 5.4 GB | ~41 tok/s | 86 | |
| 1-bit Bonsai 27B Bonsai 27B · 1-bit (binary, ~1.125 bpw) | 3.8 GB | ~47 tok/s | 86 |
What 16 GB unlocks
The strongest models that need the next tier up.
FAQ
What LLMs run well with 12 GB VRAM?
Qwen 2.5 Family, Qwen 3, DeepSeek R1, Phi-4 Family, Granite 3.0 all fit in 12 GB VRAM.
Which GPUs have 12 GB VRAM?
Intel Arc B580, NVIDIA GeForce RTX 3060 (12GB), AMD Radeon RX 7700 XT, NVIDIA GeForce RTX 5070.
All tiers at a glance
| Memory | Models fit | Best overall | Largest that fits |
|---|---|---|---|
| 4 GB | 32 | Gemma 3 4B Instruct | Gemma 3 4B Instruct · 2.4 GB |
| 6 GB | 56 | Granite 3.0 8B Instruct | Qwen 2.5 VL 7B Instruct · 5 GB |
| 8 GB | 71 | Granite 3.0 8B Instruct | Falcon 3 10B Instruct · 6.2 GB |
| 10 GB | 83 | Qwen 2.5 14B Instruct | Qwen 3 14B · 8.9 GB |
| 12 GB | 85 | Qwen 2.5 14B Instruct | StarCoder 2 15B · 9.4 GB |
| 16 GB | 95 | Qwen 2.5 14B Instruct | Mistral Small 3 (24B) · 14.5 GB |
| 20 GB | 111 | Qwen 2.5 14B Instruct | Gemma 4 31B · 18.7 GB |
| 24 GB | 131 | Qwen 2.5 Coder 32B | Qwen 3.6 35B-A3B · 21.1 GB |
| 32 GB | 132 | Qwen 2.5 Coder 32B | Qwen 3.6 35B-A3B · 21.1 GB |
| 48 GB | 138 | Qwen 2.5 72B Instruct | Qwen 2.5 VL 72B Instruct · 44.3 GB |
| 64 GB | 143 | Qwen 2.5 72B Instruct | Qwen3-Coder-Next (80B-A3B MoE) · 48.3 GB |
| 80 GB | 152 | Qwen 2.5 72B Instruct | Qwen 3.5 122B-A10B · 73.7 GB |
Buy this hardware Intel Arc B580 12GB — 12 GB VRAM · 190 W board powerDeploy in the cloud now RTX 4090 on RunPod — from $0.34/hr · rate checked 2026-07
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
Can-I-Run Pages Near 12 GB
- DeepSeek R1 on NVIDIA GeForce RTX 5070
- DeepSeek R1 on NVIDIA GeForce RTX 4070 Ti
- DeepSeek R1 on NVIDIA GeForce RTX 4070 Super
- DeepSeek R1 on NVIDIA GeForce RTX 4070
- DeepSeek R1 on NVIDIA GeForce RTX 3060 (12GB)
- DeepSeek R1 on Intel Arc B580
- Llama 3.3 on NVIDIA GeForce RTX 5070
- Llama 3.3 on NVIDIA GeForce RTX 4070 Ti