 |
VOOZH | about |
Cerebras-GPT 13B needs ~22.3 GB VRAM. RTX A4500 20GB has 20.0 GB. With Q5_K_M quantization, expect ~32 tok/s.
Operating mode
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
2.3 GB over capacity — needs offload or smaller quantization
Fit status
Very compromised (needs ~1 GB host RAM)
Decode
32.4 tok/s
TTFT
5981 ms
Safe context
12K
Memory
22.3 GB / 20.0 GB
Offload
10%
It fits through host-memory offload, and offload is the main reason performance drops.
CPU or host-memory offload is active
About 10% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.
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.
Remove offload with more accelerator memory
Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.
Buy headroom, not only minimum fit
A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.
Increase host RAM if you keep offloading
This setup may need roughly 1.0 GB of extra host RAM just for the offloaded portion, before OS and other tools.
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | B | Tight fit | 54.4 tok/s | 1941 ms | 12K |
| Coding | B | Very compromised (needs ~1 GB host RAM) | 32.4 tok/s | 5981 ms | 12K |
| Agentic Coding | F | Too heavy | 15.1 tok/s | 18674 ms | 12K |
| Reasoning | B | Very compromised (needs ~1 GB host RAM) | 32.4 tok/s | 7069 ms | 12K |
| RAG | F | Too heavy | 15.1 tok/s | 23343 ms |
How Cerebras-GPT 13B (13B params) fits at each quantization level on RTX A4500 20GB (20.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 5.1 GB | Low | B63 |
Q3_K_S | 3 | 6.4 GB | Low | B64 |
NVFP4 | 4 |
Copy-paste commands to run Cerebras-GPT 13B on your machine.
Run
docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
--hf-repo "cerebras/Cerebras-GPT-13B" \
--hf-file "Cerebras-GPT-13B-Q5_K_M.gguf" \
-c 4096 -ngl 99Upgrade options
Removes host-memory offload, which is usually the single biggest latency and throughput win.
Raises estimated decode speed by about 120%.
~$1,499 MSRP
Removes host-memory offload, which is usually the single biggest latency and throughput win.
Raises estimated decode speed by about 158%.
~$1,599 MSRP
Removes host-memory offload, which is usually the single biggest latency and throughput win.
Raises estimated decode speed by about 102%.
~$3,200 MSRP
| 12K |
7.3 GB |
| Medium |
| B65 |
Q4_K_M | 4 | 7.9 GB | Medium | B65 |
Q5_K_M | 5 | 9.4 GB | High | B66 |
Q6_K | 6 | 10.7 GB | High | B67 |
Q8_0Best for your GPU | 8 | 13.9 GB | Very High | B66 |
F16 | 16 | 26.7 GB | Maximum | F0 |
Remove offload with more accelerator memory. Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.
