Compare · GPUs
RTX 3080 vs RTX 3060 for local LLMs
It runs 16 of 39 models fully in memory, against 14 for the RTX 3080, thanks to 12 GB of VRAM versus 10 GB. The RTX 3080 is about 2.3x faster on the models both can hold, so it only wins if you stick to small models.
It runs 16 of 39 models fully in memory, against 14 for the RTX 3080, thanks to 12 GB of VRAM versus 10 GB. The RTX 3080 is about 2.3x faster on the models both can hold, so it only wins if you stick to small models.
Every model at 8K context with the best quantization that fits. Offload assumes 32 GB of system RAM. Green marks the better result. Rows where the two differ come first.
| Model | RTX 3080 | RTX 3060 |
|---|---|---|
| Gemma 4 12B | Offload~9.6 tok/sQ4_K_M | Runs~33 tok/sQ4_K_M |
| Mistral Nemo 12B | Offload~10 tok/sQ4_K_M | Runs~33 tok/sQ4_K_M |
| Qwen3 0.6B | Runs~780 tok/sQ8_0 | Runs~370 tok/sQ8_0 |
| Gemma 3 1B | Runs~470 tok/sQ8_0 | Runs~220 tok/sQ8_0 |
| Llama 3.2 1B | Runs~390 tok/sQ8_0 | Runs~180 tok/sQ8_0 |
| Llama 3.2 3B | Runs~150 tok/sQ8_0 | Runs~69 tok/sQ8_0 |
| Phi-4 Mini 3.8B | Runs~120 tok/sQ8_0 | Runs~58 tok/sQ8_0 |
| Qwen3 4B | Runs~120 tok/sQ8_0 | Runs~55 tok/sQ8_0 |
| Gemma 3 4B | Runs~110 tok/sQ8_0 | Runs~51 tok/sQ8_0 |
| Qwen3.5 4B | Runs~99 tok/sQ8_0 | Runs~47 tok/sQ8_0 |
| Mistral 7B v0.3 | Runs~84 tok/sQ6_K | Runs~31 tok/sQ8_0 |
| Qwen2.5-Coder 7B | Runs~79 tok/sQ6_K | Runs~29 tok/sQ8_0 |
| Llama 3.1 8B | Runs~75 tok/sQ6_K | Runs~28 tok/sQ8_0 |
| DeepSeek-R1 Distill 8B | Runs~75 tok/sQ6_K | Runs~28 tok/sQ8_0 |
| Qwen3 8B | Runs~74 tok/sQ6_K | Runs~27 tok/sQ8_0 |
| Qwen3.5 9B | Runs~74 tok/sQ5_K_M | Runs~30 tok/sQ6_K |
| Gemma 3 12B | Offload~9.4 tok/sQ4_K_M | Offload~4.9 tok/sQ4_K_M |
| Phi-4 14B | Offload~7.8 tok/sQ4_K_M | Offload~4.0 tok/sQ4_K_M |
| Qwen3 14B | Offload~7.7 tok/sQ4_K_M | Offload~4.0 tok/sQ4_K_M |
| DeepSeek-R1 Distill 14B | Offload~7.7 tok/sQ4_K_M | Offload~4.0 tok/sQ4_K_M |
| gpt-oss 20B | Offload~15 tok/sQ4_K_M | Offload~8.6 tok/sQ4_K_M |
| Codestral 22B | Offload~4.3 tok/sQ4_K_M | Offload~2.7 tok/sQ4_K_M |
| Mistral Small 3 24B | Offload~3.9 tok/sQ4_K_M | Offload~2.5 tok/sQ4_K_M |
| Devstral Small 24B | Offload~3.9 tok/sQ4_K_M | Offload~2.5 tok/sQ4_K_M |
| Gemma 4 26B-A4B (MoE) | Offload~13 tok/sQ4_K_M | Offload~8.7 tok/sQ4_K_M |
| Gemma 3 27B | Offload~3.1 tok/sQ4_K_M | Offload~2.2 tok/sQ4_K_M |
| Qwen3.6 27B | Offload~3.3 tok/sQ4_K_M | Offload~2.2 tok/sQ4_K_M |
| Qwen3 30B-A3B (MoE) | Offload~14 tok/sQ4_K_M | Offload~10 tok/sQ4_K_M |
| Qwen3-Coder 30B-A3B | Offload~14 tok/sQ4_K_M | Offload~10 tok/sQ4_K_M |
| Gemma 4 31B | Offload~2.7 tok/sQ4_K_M | Offload~1.9 tok/sQ4_K_M |
| Qwen3 32B | Offload~2.6 tok/sQ4_K_M | Offload~1.8 tok/sQ4_K_M |
| DeepSeek-R1 Distill 32B | Offload~2.6 tok/sQ4_K_M | Offload~1.8 tok/sQ4_K_M |
| Qwen2.5-Coder 32B | Offload~2.6 tok/sQ4_K_M | Offload~1.8 tok/sQ4_K_M |
| Qwen3.6 35B-A3B (MoE) | Offload~14 tok/sQ4_K_M | Offload~11 tok/sQ4_K_M |
| Llama 3.3 70B | Won’t fit | Won’t fit |
| DeepSeek-R1 Distill 70B | Won’t fit | Won’t fit |
| Llama 4 Scout 109B-A17B | Won’t fit | Won’t fit |
| gpt-oss 120B | Won’t fit | Won’t fit |
| Qwen3 235B-A22B (MoE) | Won’t fit | Won’t fit |
Which should you buy?
The RTX 3060 is the better card for local AI. It runs 16 of 39 models fully in memory, against 14 for the RTX 3080, thanks to 12 GB of VRAM versus 10 GB. The RTX 3080 is about 2.3x faster on the models both can hold, so it only wins if you stick to small models.
Only the RTX 3060 fits Mistral Nemo 12B and Gemma 4 12B fully in memory.
See every result on its own page: RTX 3080 and RTX 3060.
Questions
Is the RTX 3080 or the RTX 3060 better for running LLMs?
The RTX 3060 is the better card for local AI. It runs 16 of 39 models fully in memory, against 14 for the RTX 3080, thanks to 12 GB of VRAM versus 10 GB. The RTX 3080 is about 2.3x faster on the models both can hold, so it only wins if you stick to small models.
Which models can the RTX 3060 run that the RTX 3080 can't?
At 8K context, only the RTX 3060 holds Mistral Nemo 12B and Gemma 4 12B fully in VRAM.
Does the RTX 3080 or RTX 3060 need anything else to run local AI?
Only the software. Install Ollama or LM Studio, pick a model that fits in VRAM, and it runs on the GPU. 32 GB of system RAM helps when a model spills over, but that path is much slower.
More comparisons
Estimates from memory size and bandwidth at 8K context. Real speed varies with software, drivers and settings.