40 GB VRAM (+8)1555 GB/s (+327)
B
Raises estimated decode speed by about 64%.72.8 tok/s decode
Raises estimated decode speed by about 64%.
Adds memory headroom for longer context windows and future model growth.
~$10,000 MSRP
Can Aya Expanse 32B run on AMD Instinct MI100 32GB?
YES — Runs Great
B60Good
○Estimated from fit model
Aya Expanse 32B needs ~26.1 GB VRAM. AMD Instinct MI100 32GB has 32.0 GB. With Q4_K_M quantization, expect ~45 tok/s.
Runtime: llama.cppCapacity: RoomyBandwidth: HighStack: StandardBottleneck: Balanced
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) — 26.1 GB, 44.5 tok/s, Runs well
26.1 GB required32.0 GB available
82% VRAM used
Fit status
Runs well
Decode
44.5 tok/s
TTFT
4354 ms
Safe context
8K
Memory
26.1 GB / 32.0 GB
Memory breakdown
Weights19.5 GB
KV Cache2.4 GB
Runtime0.9 GB
Headroom3.2 GB
See how fast it feels
See how fast it feelsAya Expanse 32B on AMD Instinct MI100 32GB
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: 44.5 tok/s decode · 4.4s TTFT (warm) · 111 tok/s prefill
What limits this setup
This setup is broadly balanced for this model.
No major red flags
This recommendation has enough memory headroom and acceptable estimated speed for the selected workload.
Best improvement path
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | B | Runs well | 44.5 tok/s | 2375 ms | 8K |
| Coding | B | Runs well | 44.5 tok/s | 4354 ms | 8K |
| Agentic Coding | B | Tight fit | 44.5 tok/s | 6332 ms | 8K |
| Reasoning | B | Runs well | 44.5 tok/s | 5145 ms | 8K |
| RAG | B | Tight fit | 44.5 tok/s | 7916 ms | 8K |
Quantization options
How Aya Expanse 32B (32B params) fits at each quantization level on AMD Instinct MI100 32GB (32.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 12.5 GB | Low | C53 |
Q3_K_S | 3 | 15.7 GB | Low | C55 |
NVFP4 | 4 |
Get started
Copy-paste commands to run Aya Expanse 32B on your machine.
Run
ollama run aya-expanse:32bUpgrade options