32 GB VRAM (+8)1792 GB/s (+1492)
B
Raises estimated decode speed by about 931%.43.3 tok/s decode
Raises estimated decode speed by about 931%.
Adds memory headroom for longer context windows and future model growth.
~$1,999 MSRP
Can Aya Expanse 32B run on NVIDIA L4 24GB?
YES — With Offload
C50Usable
○Estimated from fit model
Aya Expanse 32B needs ~25.3 GB VRAM. NVIDIA L4 24GB has 24.0 GB. With Q4_K_M quantization, expect ~4 tok/s.
Runtime: llama.cppCapacity: OffloadBandwidth: LowStack: StandardBottleneck: Memory bandwidth
Share:
Operating mode
Choose the run profile you care about
Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
Select quantization to explore
Q4_K_M (Medium quality) — 25.3 GB, 4.2 tok/s, Runs with offload (needs ~1 GB host RAM)
25.3 GB required24.0 GB available
105% VRAM needed
1.3 GB over capacity — needs offload or smaller quantization
Fit status
Runs with offload (needs ~1 GB host RAM)
Decode
4.2 tok/s
TTFT
45948 ms
Safe context
8K
Memory
25.3 GB / 24.0 GB
Memory breakdown
Weights19.5 GB
KV Cache2.4 GB
Runtime0.9 GB
Headroom2.4 GB
See how fast it feels
See how fast it feelsAya Expanse 32B on NVIDIA L4 24GB
1st promptCold start — includes initialization
>What is local AI inference?
Local AI inference means running an AI language model directly on your own hardware — your laptop, desktop, or server — instead of sending requests to a remote cloud API.
When you run inference locally the model weights are loaded into your GPU or unified memory. Each token you generate requires reading those weights from memory, so memory bandwidth is the main bottleneck for decode speed.
Key benefits of running locally:
- Full privacy: your prompts never leave your machine
- No per-token cost or rate limits
- Works offline once the model is downloaded
- Latency depends only on your hardware
2nd promptWarm — model ready, no init overhead
>How much VRAM do I need?
It depends on the model size and quantization level. A rough rule of thumb:
Model size Q4 (4-bit) Q8 (8-bit) FP16
7B params ~4.3 GB ~7.5 GB ~14 GB
13B params ~7.9 GB ~13.9 GB ~26 GB
70B params ~42.7 GB ~74.9 GB ~140 GB
Most people use 4-bit quantization (Q4_K_M) which gives 90-95% of full quality at a fraction of the memory. A 24 GB GPU can comfortably run most 7B-13B models.
Estimated: 4.2 tok/s decode · 45.9s TTFT (warm) · 11 tok/s prefill
What limits this setup
This model fits, but memory bandwidth is the part holding decode speed back.
Throughput will feel slow
Estimated decode speed is only 4.2 tok/s, so this is more of a technical fit than a comfortable daily-driver setup.
Very little memory headroom
You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.
Best improvement path
Prioritize bandwidth, not only capacity
If this workload feels slow, the next useful step is often a GPU tier with materially faster memory bandwidth rather than only a small bump in capacity.
Buy headroom, not only minimum fit
A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | C | Runs with offload (needs ~0 GB host RAM) | 4.7 tok/s | 22581 ms | 8K |
| Coding | C | Runs with offload (needs ~1 GB host RAM) | 4.2 tok/s | 45948 ms | 8K |
| Agentic Coding | D | Very compromised (needs ~2.6 GB host RAM) | 3.5 tok/s | 81158 ms | 8K |
| Reasoning | C | Runs with offload (needs ~1 GB host RAM) | 4.2 tok/s | 54302 ms | 8K |
| RAG | D | Very compromised (needs ~2.6 GB host RAM) | 3.5 tok/s | 101448 ms | 8K |
Quantization options
How Aya Expanse 32B (32B params) fits at each quantization level on NVIDIA L4 24GB (24.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 12.5 GB | Low | B56 |
Q3_K_S | 3 | 15.7 GB | Low | B55 |
NVFP4Best for your GPU | 4 | 17.9 GB | Medium | C55 |
Q4_K_M | 4 | 19.5 GB | Medium | F0 |
Q5_K_M | 5 | 23.0 GB | High | F0 |
Q6_K | 6 | 26.2 GB | High | F0 |
Q8_0 | 8 | 34.2 GB | Very High | F0 |
F16 | 16 | 65.6 GB | Maximum | F0 |
Get started
Copy-paste commands to run Aya Expanse 32B on your machine.
Run
ollama run aya-expanse:32bUpgrade options
Hardware that runs Aya Expanse 32B well
32 GB VRAM (+8)896 GB/s (+596)
B
Raises estimated decode speed by about 898%.41.9 tok/s decode
Raises estimated decode speed by about 898%.
Adds memory headroom for longer context windows and future model growth.
~$2,499 MSRP
32 GB VRAM (+8)576 GB/s (+276)
B
Raises estimated decode speed by about 512%.25.7 tok/s decode
Raises estimated decode speed by about 512%.
Adds memory headroom for longer context windows and future model growth.
~$4,000 MSRP