Can you run Gemma 4 E2B-it on GTX 1660 Ti?
Short answer
Yes, but only just — Gemma 4 E2B-it leaves 1.7 GB spare on GTX 1660 Ti.
- Model file
- 3.2 GB
- Needs in VRAM, with context
- about 4.3 GB
- Card memory
- 6 GB
- Measured on this hardware
- not measured
How this was worked out
Gemma 4 E2B-it is a 3.2 GB file. Running it needs the weights in memory plus room for the context window and the runtime, which comes to roughly 4.3 GB in practice. GTX 1660 Ti has 6 GB, which leaves 1.7 GB. That is arithmetic rather than a benchmark: it tells you whether the model loads, not how it feels.
What we have actually measured
We have not timed Gemma 4 E2B-it on GTX 1660 Ti. The closest figure we have is 69 tokens per second on RTX 3060 12 GB, Vulkan backend. Rather than scale that number to your card and present the result as a benchmark, this page leaves it out. The method is published at /benchmarks if you would like to run it yourself.
If it is tight
1.7 GB is not much of a margin. Your desktop compositor and anything with a video playing are competing for the same memory, so this one loads but can fall back to the CPU under pressure. If it stutters, drop to the next model down. Nightjar sizes the context window to the memory it finds, so a smaller card gets a shorter window rather than a failed load.
Smaller models that fit the same card
- Llama 3.2 1B on GTX 1660 Ti — 0.75 GB
- Qwen 3.5 2B on GTX 1660 Ti — 1.2 GB
- Qwen 3.5 4B on GTX 1660 Ti — 2.6 GB
Related
- Up a levelCan your PC run a local AI model?
- Closely relatedCan you run Gemma 4 12B-it on GTX 1660 Ti?
- Closely relatedCan you run Gemma 4 26B-A4B-it on GTX 1660 Ti?
- Closely relatedCan you run Gemma 4 31B-it on GTX 1660 Ti?
- Closely relatedCan you run Gemma 4 E2B-it on Intel Arc A770?
- Elsewhere on the siteLocal AI on GTX 1660 Ti: which models run and how fast
- Elsewhere on the siteGemma 4 E2B-it in Nightjar: speed, size and what it is good at