What does OscarJHernandez run on a GMKtec EVO-X2?

OscarJHernandez (self-reported, 27 Oct 2026): I've been experimenting with different models and quantizations, primarily on the GMKtec EVO-X2. The benchmark repo is https://github.com/OscarJHernandez/gmktec-evo-x2-local-llm-benchmarks. Community measurement, not ours.

Is the GMKtec EVO-X2 benchmark finished?

Not yet. OscarJHernandez wrote that it took several days to run since he collected data over many models, and that he may share more detailed findings once he has finished analyzing the experiments. Self-reported, community measurement, not ours.

Can I test open models in the browser before buying a GMKtec EVO-X2?

Yes. The free browser test is at https://localllmrig.com/.

OscarJHernandez's GMKtec EVO-X2 128 GB: models, quants, and a benchmark still being analyzed

OscarJHernandez experiments with models and quantizations on a GMKtec EVO-X2 with 128 GB of unified memory. He is not planning an immediate upgrade. The multi-day run across many models is still being analyzed. The quotes below are self-reported, community measurement, not ours. Repo: https://github.com/OscarJHernandez/gmktec-evo-x2-local-llm-benchmarks.

OscarJHernandez, quoted by GitHub username with written permission. 27 Oct 2026. Self-reported. Not a localLLMrig measurement.

Test Gemma 4 31B, Recursive Flint and Nemotron 3.5 Super VL side by side in your browser, free.

github.com/OscarJHernandez. github.com/OscarJHernandez/gmktec-evo-x2-local-llm-benchmarks.

What are you running on the GMKtec EVO-X2?

I've been experimenting with different models and quantizations, primarily on the GMKtec EVO-X2.

Are you planning to upgrade that hardware?

I'm not considering an immediate upgrade. I'm more interested in understanding the limitations of my current hardware.

New model hardware requirements and price updates by email. We email a link to confirm the address. Nothing is sent until you click it. Privacy notice.

Where do you look for memory and speed numbers?

Mostly documentation and community benchmarks. Reliable, comparable measurements are often difficult to find.

What would make you trust a benchmark?

Clear methodology, reproducible experiments, and multiple measurements would make me trust the results.

What should a hardware page tell you up front?

Realistic performance estimates, memory requirements, and hardware compatibility would be particularly useful.

The multi-day run

I may be able to share more detailed findings once I've finished analyzing my experiments.

I'd be happy to pass along my learnings once I have reviewed the data I collected from run. It took several days to run since I collected data over many models.

Repo: github.com/OscarJHernandez/gmktec-evo-x2-local-llm-benchmarks.

On our list: GMKtec EVO-X2 128GB, What fits in 128 GB, Mac vs PC for a local LLM.

The numbers on this page were measured by the owner on that hardware. They are not a localLLMrig benchmark, and they do not change the fit scores in the calculator.

Questions

What does OscarJHernandez run on a GMKtec EVO-X2?: OscarJHernandez (self-reported, 27 Oct 2026): I've been experimenting with different models and quantizations, primarily on the GMKtec EVO-X2. The benchmark repo is https://github.com/OscarJHernandez/gmktec-evo-x2-local-llm-benchmarks. Community measurement, not ours.
Is the GMKtec EVO-X2 benchmark finished?: Not yet. OscarJHernandez wrote that it took several days to run since he collected data over many models, and that he may share more detailed findings once he has finished analyzing the experiments. Self-reported, community measurement, not ours.
Can I test open models in the browser before buying a GMKtec EVO-X2?: Yes. The free browser test is at https://localllmrig.com/.

Find hardware · All builder interviews · Send us your llama-bench result