NVIDIA L4 for local LLMs
Written by Jakub Rusinowski · Last updated
With 24 GB of GDDR6 at 300 GB/s, the L4 runs 109 catalogued models at Q4_K_M with 8K context. The largest that fits is Qwen 3 32B (~22.8 GB), and the top pick is Laguna XS 2.1 33B-A3B at 44–85 tok/s.
24 GB of GDDR6 at 300 GB/s in a 72 W, single-slot, low-profile card. Plenty of memory, little bandwidth. Common in cloud instances.
Models that run on the L4
Q4_K_M, 8K context, 24 GB usable. Ranked by quality and speed.
| Model | Memory · marker = 24 GB | VRAM | Speed |
|---|---|---|---|
| Laguna XS 2.1 33B-A3B Poolside Laguna XS 2.1 | 20.7 GB | 44–85 tok/s | |
| Granite 4.0 Small-H 32B-A9B IBM Granite 4.0 | 20.1 GB | 24–46 tok/s | |
| Nemotron-Cascade 2 30B-A3B Nemotron Cascade 2 | 19.9 GB | 42–80 tok/s | |
| Qwen 3 30B-A3B (MoE) Qwen 3 | 20 GB | 56–107 tok/s | |
| Qwen3-Coder 30B-A3B (MoE) Qwen3-Coder | 20 GB | 56–107 tok/s | |
| GLM-4.7-Flash 30B-A3B GLM-4.7 / GLM-Z1 | 18.9 GB | 45–87 tok/s | |
| Nemotron 3 Nano Omni 30B-A3B Nemotron 3 Nano Omni | 18.9 GB | 45–87 tok/s | |
| North Mini Code 1.0 30B-A3B North Mini Code | 18.9 GB | 45–87 tok/s | |
| Nemotron 3.5 Lightning 30B-A3B Nemotron 3.5 | 18.9 GB | 45–87 tok/s | |
| Trinity Mini Trinity | 16.9 GB | 74–142 tok/s | |
| Gemma 4 26B-A4B Gemma 4 | 16.5 GB | 40–77 tok/s | |
| GPT-OSS 20B GPT-OSS | 13.8 GB | 59–114 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 2026-10-07.
- Memory
- 24 GB GDDR6
- Memory bandwidth
- 300 GB/s
- Memory bus
- 192-bit
- Architecture
- Ada Lovelace AD104
- Series
- Data center
- Board power
- 72 W
- Release year
- 2023
- Compute backends
- CUDA
- Usable for models
- 24 GB
Similar GPUs
Frequently asked questions
Can the NVIDIA L4 run local LLMs?
Yes. With 24 GB (24 GB usable by a model) it runs 109 of the catalogued models at Q4_K_M with 8K context; the largest is Qwen 3 32B, needing about 22.8 GB.
How fast is the NVIDIA L4 for AI inference?
It is estimated to run Llama 3.1 8B at 32–61 tok/s at Q4_K_M. Llama 3.3 70B does not fit: it needs about 44 GB against 24 GB usable. These are modelled estimates from memory bandwidth, not measurements; the methodology page shows the formula.
What LLMs can I run on 24 GB?
Among the best that fit: Laguna XS 2.1 33B-A3B, Granite 4.0 Small-H 32B-A9B, Nemotron-Cascade 2 30B-A3B, Qwen 3 30B-A3B (MoE), Qwen3-Coder 30B-A3B (MoE).