Hardware

GeForce GTX 1660 Ti

GeForce GTX 1660 Ti has 6 GB GDDR6 and 288 GB/s. See which open models fit, at which quant, and roughly how fast. Figures below use the default memory (6 GB), 32 GB system RAM, and DDR5-5600 for anything that spills. Updated 2026-10-24.

Memory6 GB GDDR6

Bandwidth288 GB/s

Kindgpu

Year2019

192-bit GDDR6 at 12 Gbps = 288 GB/s. Source.

RSS for new models on this device

Models that fit

Model	Verdict	Quant	Context	tok/s
Qwen2.5 VL 7B Instruct 7.6B	Tight fit	UD-Q4_K_XL	4K	32–43
gemma 4 E4B it qat 7.5B	Runs	UD-Q4_K_XL	8K	32–44
Qwen2.5 7B Instruct 7.6B	Tight fit	Q4_K_M	4K	32–44
DeepSeek R1 Distill Qwen 7B 7.6B	Tight fit	Q4_K_M	4K	32–44
Qwen2.5 Coder 7B Instruct 128K 7.6B	Tight fit	Q4_K_M	4K	32–44
gemma 4 E4B it 7.5B	Tight fit	Q4_K_M	4K	30–41
DeepSeek R1 Distill Llama 8B 8.0B	Tight fit	IQ4_XS	4K	32–43
Llama 3.1 8B Instruct 8.0B	Tight fit	IQ4_XS	4K	32–43
Apertus 8B Instruct 2509 8.1B	Tight fit	IQ4_XS	4K	32–43
MiMo V2.6 Distill Qwen 9B 9.0B	Tight fit	Q3_K_L	8K	32–44
gemma 4 12b it 11.9B	Tight fit	UD-IQ3_XXS	2K	29–40
Qwen3.5 9B 9.0B	Tight fit	Q3_K_M	8K	32–44
Ministral 3 8B Instruct 2512 8.5B	Tight fit	UD-Q3_K_XL	4K	32–43
Ministral 3 8B Reasoning 2512 8.5B	Tight fit	UD-Q3_K_XL	4K	32–43
granite 4.2 8b 8.8B	Tight fit	Q3_K_M	4K	32–43
Qwen3 VL 8B Instruct 8.2B	Tight fit	UD-Q3_K_XL	4K	32–44
DeepSeek R1 0528 Qwen3 8B 8.2B	Tight fit	UD-Q3_K_XL	4K	32–44
Qwen3 8B 8.2B	Tight fit	UD-Q3_K_XL	4K	32–44
Qwen3 VL 8B Thinking 8.2B	Tight fit	UD-Q3_K_XL	4K	32–44
LFM2.5 8B A1B 8.5B	Tight fit	UD-IQ4_XS	8K	175–238
gemma 4 E2B it 4.6B	Tight fit	Q8_0	8K	30–40
Qwen3 14B 14.8B	Tight fit	UD-IQ1_M	4K	33–46
NVIDIA Nemotron 3.5 Bolt E4B 7.5B	Runs	UD-Q3_K_XL	8K	32–44
GLM 4.6V Flash 9.4B	Tight fit	Q3_K_S	8K	32–44
GLM Z1 9B 0414 9.4B	Tight fit	Q3_K_S	8K	32–44
GLM 4 9B 0414 9.4B	Tight fit	Q3_K_S	8K	32–44
Qwen3.8 4B Distill 4.3B	Tight fit	Q8_0	8K	32–44
Qwen3.5 4B 4.2B	Tight fit	Q8_0	8K	33–45
Ministral 3 14B Instruct 2512 13.5B	Tight fit	UD-IQ2_XXS	4K	33–46
Ministral 3 14B Reasoning 2512 13.5B	Tight fit	UD-IQ2_XXS	4K	33–46
Sharp Spark X2.5 4B 4.1B	Runs	Q6_K	8K	33–45
Qwen3 4B 4.0B	Tight fit	Q8_0	4K	32–44
Qwen3 VL 4B Instruct 4.0B	Tight fit	Q8_0	4K	32–44
Qwen3 4B Thinking 2507 4.0B	Tight fit	Q8_0	4K	32–44
Qwen3 VL 4B Thinking 4.0B	Tight fit	Q8_0	4K	32–44
NVIDIA Nemotron 3 Nano 4B 4.0B	Tight fit	Q8_0	4K	32–44
gemma 3 4b it 3.9B	Tight fit	Q8_0	8K	30–41
medgemma 4b it 3.9B	Tight fit	Q8_0	8K	30–41
Phi 4 mini instruct 3.8B	Tight fit	Q8_0	4K	34–47
Phi 4 mini reasoning 3.8B	Tight fit	Q8_0	4K	34–47

105 more models in the catalog do not fit this default profile, or ranked below the first 40 that do.

Example: Tight fit fully on your GeForce GTX 1660 Ti at UD-Q4_K_XL, 4K context, about 32–43 tok/s. Weights 4.46 GB plus 0.22 GB of KV cache — 5.02 GB of 5.22 GB usable. Weights shown in the verdict use GiB (the same 1024³ counting as GPU memory).