The test machine
One workstation, deliberately ordinary and a little old, because that is what most people asking “can my computer run this?” actually own.
| CPU | Intel Xeon E5-2665 (Sandy Bridge, 2012), 8 cores and 16 threads. It has AVX but not AVX2, so CPU results are close to a floor for a desktop. |
|---|---|
| Memory | 62 GB system RAM |
| GPU | NVIDIA GeForce RTX 3060, 12 GB, driver 580.178.04 |
| Software | Ubuntu 24.04.5 LTS, kernel 6.8, Ollama 0.11.4 (the version is stated on every result) |
| Background load | A normal desktop session. During the September 24, 2026 run about 1.2 GB (1,101 MiB) of the GPU’s memory was already in use by the desktop and a Stable Diffusion web UI, which is typical of a real machine and is reported with the results. |
The benchmark method
- Every model gets the same prompt: “Explain in about 150 words why someone might run a language model locally.”
- Temperature 0, a fixed seed and answers capped at 200 tokens. The seed only fixes the wording; it doesn’t change speed.
- One cold run with the model unloaded (this includes loading it from disk), then three warm runs. The published speed is the median warm generation rate.
- CPU-only runs set Ollama’s
num_gpuoption to 0, which puts no model layers on the GPU. The model is unloaded between CPU and GPU runs. - Memory is what Ollama itself reports while the model is loaded, with its default 4,096-token context. Longer contexts use more.
The script is bench.py, one file with no dependencies beyond Python and Ollama, so anyone can repeat the test and compare.
What we publish
- The hardware, operating system, driver and Ollama version behind every number
- The raw results as CSV, under CC BY 4.0, in the local-ai-benchmarks repository
- Real terminal output, copied from the test machine, when a page shows commands
- The date each page was tested and last updated, and a list of what changed
- What we did not test, next to what we did
What we don’t do
- Present someone else’s benchmark as our own, or round a number in a product’s favor
- Accept payment for placement or for a result
- Show an illustration as if it were a screenshot. Illustrations are labelled.
- Publish AI-generated artwork as Lorelei’s own work
Disclosures
Arynwood MCP, Cutroom and Arynwood Community are Arynwood’s own products, and pages that mention them say so. Arynwood sells local AI setup services. The site footer contains one affiliate link, for VPS hosting, marked as an affiliate link.
Corrections and updates
Found a mistake, or got a very different number on similar hardware? Email [email protected] or submit your results. Corrections are made on the page, dated, and listed in its update notes. Pages are re-tested when a new Ollama release changes the results, and at least every quarter.