# Can I run Llama 3.1 8B on an RTX 4090? | What Can I Run AI?

> Yes. Llama 3.1 8B at FP16 needs about 19.5 GB; RTX 4090 has 24.0 GB usable. Estimated tokens/sec and context headroom.

Source: https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-4090/

[Home](https://whatcanirunai.vercel.app/) / [Can I run it?](https://whatcanirunai.vercel.app/can-i-run/) / [Llama 3.1 8B](https://whatcanirunai.vercel.app/models/llama-3.1-8b-instruct/) / Llama 3.1 8B on an RTX 4090

**Yes.** Llama 3.1 8B at FP16 needs about 19.5 GB, and the GeForce RTX 4090 has 24.0 GB usable at a \~4k context.

# Can I run Llama 3.1 8B on an RTX 4090?

Llama 3.1 8B at FP16 needs about 19.5 GB, and the GeForce RTX 4090 has 24.0 GB usable. That figure is 24 GB of VRAM on the spec sheet, of which 24.0 GB is usable for weights and cache, at 1008 GB/s memory bandwidth. Meta’s widely used 8B instruct model, the default small local chat model for a few years running. These are estimates from [our formula](https://whatcanirunai.vercel.app/method/), not measured benchmarks.

Quant that fits

FP16 Fits with headroom

Memory needed

19.5 GB at FP16, \~4k context

Usable VRAM

24.0 GB of 24 GB VRAM

Est. tokens/sec

38–53 estimate

[Llama-3.1-8B-Instruct](https://whatcanirunai.vercel.app/models/llama-3.1-8b-instruct/) · [GeForce RTX 4090](https://whatcanirunai.vercel.app/hardware/rtx-4090/) · [Everything this machine runs](https://whatcanirunai.vercel.app/what-can-i-run/rtx-4090/) · [Q4 vs Q8 vs FP16](https://whatcanirunai.vercel.app/quantization/)

## Memory at each quant

| Quant    | Need vs 24.0 GB | Verdict  |
| -------- | --------------- | -------- |
| Q4\_K\_M | 6.6 GB          | Q4Fits   |
| Q8\_0    | 10.8 GB         | Q8Fits   |
| FP16     | 19.5 GB         | FP16Fits |

## Estimated tokens per second

Estimate: **38–53 tokens/sec** at FP16\. This is a bandwidth-derived range, not a bench. Efficiency is a 60–85% GPU band from [How we calculate](https://whatcanirunai.vercel.app/method/).

## Context headroom

At FP16, this machine has headroom out to about **32k tokens** under our KV estimate. Longer chats need more memory for the cache. Figures assume an FP16 KV cache.

__Context headroom for Llama 3.1 8B on an RTX 4090 at FP16__
| Context | KV cache (FP16) | Total at FP16 | Verdict   |
| ------- | --------------- | ------------- | --------- |
| 4k      | 0.5 GB          | 19.5 GB       | Fits      |
| 8k      | 1.1 GB          | 20.0 GB       | Fits      |
| 16k     | 2.1 GB          | 21.1 GB       | Fits      |
| 32k     | 4.3 GB          | 23.2 GB       | Tight     |
| 64k     | 8.6 GB          | 27.5 GB       | Won’t fit |
| 128k    | 17.2 GB         | 36.1 GB       | Won’t fit |

From the model’s config. Try other sizes in the [VRAM calculator](https://whatcanirunai.vercel.app/vram-calculator/) or see the [context length guide](https://whatcanirunai.vercel.app/context-length/).

## What to do instead

* See the [best GPU for this model family](https://whatcanirunai.vercel.app/best-gpu-for/llama/) and the [Llama-3.1-8B-Instruct model page](https://whatcanirunai.vercel.app/models/llama-3.1-8b-instruct/).

## Other hardware for Llama 3.1 8B

* [Llama 3.1 8B on an RTX 3060 12GB](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-3060-12gb/)Tight
* [Llama 3.1 8B on an RTX 3070 Ti](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-3070-ti/)Fits
* [Llama 3.1 8B on an RTX 3080](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-3080-10gb/)Fits
* [Llama 3.1 8B on an RTX 4060 Ti 16GB](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-4060-ti-16gb/)Fits
* [Llama 3.1 8B on an RTX 4070 Ti Super](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-4070-ti-super/)Fits
* [Llama 3.1 8B on an RTX 5070 Ti](https://whatcanirunai.vercel.app/can-i-run/llama-3.1-8b-instruct/rtx-5070-ti/)Fits

## Other models on an RTX 4090

* [Qwen3-32B on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/qwen3-32b/rtx-4090/)Tight
* [Qwen3-30B-A3B on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/qwen3-30b-a3b/rtx-4090/)Tight
* [Qwen3.8-27B on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/qwen3.8-27b/rtx-4090/)Fits
* [Gemma 3 27B on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/gemma-3-27b-it/rtx-4090/)Fits
* [Gemma 4 31B on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/gemma-4-31b-it/rtx-4090/)Tight
* [gpt-oss-20b on an RTX 4090](https://whatcanirunai.vercel.app/can-i-run/gpt-oss-20b/rtx-4090/)Fits

## Quick answers

Can I run Llama 3.1 8B on an RTX 4090?

Yes. Llama 3.1 8B at FP16 needs about 19.5 GB, and the GeForce RTX 4090 has 24.0 GB usable.

How much VRAM does Llama 3.1 8B need?

About 6.6 GB at Q4\_K\_M, 10.8 GB at Q8\_0 and 19.5 GB at FP16, estimated for a \~4k context.

How fast is Llama 3.1 8B on an RTX 4090?

Our estimate is 38–53 tokens per second at FP16\. It is not a measured benchmark.

What context length can I use?

At FP16, our KV estimate supports up to about 32k tokens before the cache pushes the total over usable memory.

---
Agent guide: https://whatcanirunai.vercel.app/llms.txt · JSON API: https://whatcanirunai.vercel.app/developers/