NVIDIA GeForce RTX 3080 (10GB) for local LLMs
Written by Jakub Rusinowski · Last updated
With 10 GB at 760 GB/s, the RTX 3080 (10GB) runs 62 catalogued models at Q4_K_M with 8K context. The largest that fits is OLMo 2 7B Instruct (~9.5 GB), and the top pick is Gemma 4 12B (Unified) at 34–65 tok/s.
10 GB VRAM limits to 8B models but with very fast bandwidth. The 12GB variant (RTX 3080 Ti / 3080 12GB) is significantly better for AI.
Models that run on the RTX 3080 (10GB)
Q4_K_M, 8K context, 10 GB usable. Ranked by quality and speed.
| Model | Memory · marker = 10 GB | VRAM | Speed |
|---|---|---|---|
| Gemma 4 12B (Unified) Gemma 4 | 8 GB | 34–65 tok/s | |
| Bielik PL 11B v3.0 Instruct Bielik | 7.4 GB | 36–69 tok/s | |
| Llama 3.2 Vision 11B Llama 3.2 Vision | 8.5 GB | 37–71 tok/s | |
| Llama 3.2 11B Vision Instruct Llama 3.2 Family | 8.5 GB | 37–71 tok/s | |
| Falcon 3 10B Instruct Falcon 3 | 8.4 GB | 38–72 tok/s | |
| Qwen 3.5 9B Qwen 3.5 | 6.5 GB | 48–92 tok/s | |
| GLM-4 9B GLM-4.7 / GLM-Z1 | 6.2 GB | 42–81 tok/s | |
| K2 Horizon 7B K2 Horizon | 7.4 GB | 42–81 tok/s | |
| GLM-4.6V-Flash 9B GLM-4.6V | 6.2 GB | 42–81 tok/s | |
| EuroLLM 9B EuroLLM | 6.2 GB | 42–81 tok/s | |
| InternLM 3 8B Instruct InternLM 3 | 6.5 GB | 48–92 tok/s | |
| LFM2.5-8B-A1B LFM2.5 | 5.8 GB | 103–198 tok/s |
Buy it or rent the same memory
Buy the card, or rent a GPU with the same memory by the hour to try models first.
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.
Speed vs other GPUs
Llama 3.1 8B, Q4_K_M. Estimated ranges. How this is calculated
Specifications
Specs last updated 2026-07-12.
- Memory
- 10 GB
- Memory bandwidth
- 760 GB/s
- Architecture
- Ampere GA102
- Series
- RTX 30-series
- Board power
- 320 W
- Release year
- 2020
- Launch price
- $699
- Usable for models
- 10 GB
Similar GPUs
Frequently asked questions
Can the NVIDIA GeForce RTX 3080 (10GB) run local LLMs?
Yes. With 10 GB (10 GB usable by a model) it runs 62 of the catalogued models at Q4_K_M with 8K context; the largest is OLMo 2 7B Instruct, needing about 9.5 GB.
How fast is the NVIDIA GeForce RTX 3080 (10GB) for AI inference?
It is estimated to run Llama 3.1 8B at 50–97 tok/s at Q4_K_M. Llama 3.3 70B does not fit: it needs about 44 GB against 10 GB usable. These are modelled estimates from memory bandwidth, not measurements; the methodology page shows the formula.
What LLMs can I run on 10 GB?
Among the best that fit: Gemma 4 12B (Unified), Bielik PL 11B v3.0 Instruct, Llama 3.2 Vision 11B, Llama 3.2 11B Vision Instruct, Falcon 3 10B Instruct. The quickest start is Ollama: ollama run llama3.1:8b.
Keep going
Can I run it on the RTX 3080 (10GB)?
More for this card
Tools