NVIDIA GeForce GTX 1660
No editorial image yet — generic vendor mark shown. Credentials in spec table below.
Turing mid-range without RT cores. 6 GB VRAM fits 7B Q4 with short context. No Tensor cores or FP16 acceleration on consumer Turing-LITE, so inference is bandwidth-limited at ~20-30 tok/s for 7B. Common in pre-built budget gaming rigs.
Affiliate disclosure: as an Amazon Associate and partner of other retailers, we earn from qualifying purchases. The verdict on this page is our editorial opinion; affiliate links never influence what we recommend.
Sub-scores sum to 312 / 1000. Headline = 312 × 0.70 (Estimated-confidence discount) = 218. This is an algorithmic performance-tier score — distinct from, and often lower than, the editorial “Our verdict” below, which weighs value and real-world fit (especially for hardware we haven’t measured yet). How scoring works →
Extrapolated from 192 GB/s bandwidth — 23.0 tok/s estimated. No measured benchmarks yet.
Plain-English: Edge-of-fit for 7B; expect compromises.
Verdicts extrapolated from catalog VRAM + bandwidth + ecosystem flags. Hover any chip for the rationale. Want measured numbers? Submit your own run with runlocalai-bench --submit.
This card is for the operator who already owns a GTX 1660 from a gaming build and wants to dip into local AI without spending a dime on hardware. It runs 7B Q4 models at 20-30 tok/s, usable for chat but not real-time interaction. 6 GB VRAM fits 7B Q4 with short context (2K tokens), but any larger model or longer context will spill to system RAM and tank performance. The lack of Tensor cores and FP16 acceleration means inference is purely bandwidth-limited, so don't expect speed gains from quantization beyond Q4. Pass on this card if you're buying new for AI; a used RTX 3060 12 GB costs ~$180 and triples VRAM with better software support. At $130 used, it's a stopgap for budget tinkerers, not a primary inference card.
›Why this rating
The GTX 1660 is usable for small models but severely limited by 6 GB VRAM and no Tensor cores. It scores low because most local AI workloads benefit from more memory and acceleration, making this a niche budget option.
Overview
Turing mid-range without RT cores. 6 GB VRAM fits 7B Q4 with short context. No Tensor cores or FP16 acceleration on consumer Turing-LITE, so inference is bandwidth-limited at ~20-30 tok/s for 7B. Common in pre-built budget gaming rigs.
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Specs
| VRAM | 6 GB |
| Power draw (peak) | 120 W |
| Released | 2019 |
| MSRP | $219 |
| Backends | CUDA Vulkan |
Models that fit
Open-weight models small enough to run on NVIDIA GeForce GTX 1660 with usable context.
Hardware worth comparing
The closest alternatives by price, memory bandwidth, and form factor, plus a step up and down — so you can frame the buying decision against real options.
Frequently asked
What models can NVIDIA GeForce GTX 1660 run?
Does NVIDIA GeForce GTX 1660 support CUDA?
How much does NVIDIA GeForce GTX 1660 cost?
Where next?
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify hardware specifications.