Question 1

Can I run DeepSeek R1 Distill Llama 8B on a NVIDIA GeForce GTX 1660 Super?

Accepted Answer

Mostly — NVIDIA GeForce GTX 1660 Super runs DeepSeek R1 Distill Llama 8B (Q4_K_M) with partial GPU offload. Expect slower speeds than a fully fitting card. Partial GPU offload — 6 GB VRAM is above the 5.5 GB minimum but below the 8 GB recommendation. Some layers will spill to RAM.

Question 2

What quantization of DeepSeek R1 Distill Llama 8B should I use on a NVIDIA GeForce GTX 1660 Super?

Accepted Answer

For 6 GB VRAM on the NVIDIA GeForce GTX 1660 Super, the Q4_K_M variant is the best fit. Estimated ~38.4 tokens/sec on the Q4_K_M quantization.

Question 3

How fast does DeepSeek R1 Distill Llama 8B run on NVIDIA GeForce GTX 1660 Super?

Accepted Answer

Roughly 38.4 tokens/sec for Q4_K_M. Real speed depends on context length, backend (Ollama, llama.cpp, LM Studio), and KV cache size.

Question 4

What if NVIDIA GeForce GTX 1660 Super is not enough for DeepSeek R1 Distill Llama 8B?

Accepted Answer

Try a smaller quantization variant or run on CPU + RAM with the trade-off of slower generation.

Quantization	File Size	Min VRAM	Rec VRAM	Context	Verdict	Estimated tok/s
Q4_K_MBest fit	4.9 GB	5.5 GB	8 GB	8K / 128K	Partial GPU	~38.4
Q8_0	8.5 GB	9.5 GB	12 GB	8K / 128K	Hybrid CPU+GPU	~13

Can I Run DeepSeek R1 Distill Llama 8B on NVIDIA GeForce GTX 1660 Super?

Share this matchup

Every DeepSeek R1 Distill Llama 8B quantization on NVIDIA GeForce GTX 1660 Super