llmrun

NVIDIA · Turing

Best AI Models for NVIDIA GeForce GTX 1660 (6 GB)

Runs only small models well; 7–9B only just fits.

6 GB GDDR5192 GB/s10 TFLOPS FP16120 WTuring 2019
Buyabout $219

6 GB at 192 GB/s: ~7B at Q4 is tight. No Tensor Cores.

What LLMs can the GTX 1660 run?

#ModelSpeedllmrun ScoreQuant · VRAM
1
Qwen3.5 4BAlibabaHow to run →
Q4_K_M · 3.4 GB · 38.3 t/s
38.3 tok/sQ4_K_M · 3.4 GB
2
Gemma 3 4B ITGoogle
Q4_K_M · 5.0 GB · 43.9 t/s
43.9 tok/sQ4_K_M · 5.0 GB
3
Gemma 4 E2B ITGoogle
Q4_K_M · 3.5 GB · 36.4 t/s
36.4 tok/sQ4_K_M · 3.5 GB
4
Spark X2.6 4BXHToken
Q4_K_M · 4.3 GB · 42.2 t/s
42.2 tok/sQ4_K_M · 4.3 GB
5
MiniCPM5 2BOpenBMB
Q4_K_M · 2.5 GB · 65.7 t/s
65.7 tok/sQ4_K_M · 2.5 GB
See all 193 models on the GTX 1660 →Best quant that fits · speeds estimated on this card

About the GTX 1660

The GTX 1660 is an entry card in the 6 GB class: it runs only small models well. It is power-hungry for its speed, and the best speed per dollar of the cards compared at ≈$105 used. Turing generation, released 2019.

Speed in 6 GB#4192 GB/s
Per watt#50.19 tok/s per W on Qwen3.5 9B (9B params)
Per dollar#121.5 tok/s per $100 on Qwen3.5 9B (9B params)
GenerationTuringreleased 2019

GTX 1660 strengths and weaknesses

Strengths

  • Runs 7–9B models: Qwen3.5 9B at 22.6 tok/s in 5.8 GB
  • Low power draw: 120 W board power; a 250 W PSU is enough

Weaknesses

  • 14B models don't fit: Qwen3 14B needs 11.9 GB at Q4 vs 6 GB here
  • Little room for context: Qwen3.5 9B uses 5.8 of 6 GB, 0.2 GB left
  • Slow prompt reading: A 4K-token prompt on Qwen3.5 4B waits 8.4 s for the first token
  • No FP8 or NVFP4: Turing predates both; use GGUF, AWQ or GPTQ builds

Speed per dollar and per watt

On Qwen3.5 9B (9B params) at Q4_K_M. New prices are compared with new, used with used. The last two columns are the speed you get per $100 of price and per 100 W of board power; higher is better. The GTX 1660 has no new price, so the New view shows it at its MSRP.

CardNew pricetok/stok/s per $100tok/s per 100 W
GTX 16606 GB
about $21922.6
10.3
18.8
RTX 50608 GBBest new value
about $46045.8
10.0
31.6
RTX 507012 GB
about $84068.7
8.2
27.5
RTX 40608 GB
about $52027.8
5.3
24.2
CardUsed pricetok/stok/s per $100tok/s per 100 W
GTX 16606 GB
about $10522.6
21.5
18.8
RTX 3060 12GB12 GB
about $32936.8
11.2
21.6
RTX 50608 GBBest used value
about $43045.8
10.7
31.6
RTX 407012 GB
about $57551.5
9.0
25.8
RTX 40608 GB
about $33527.8
8.3
24.2
RTX 4070 SUPER12 GB
about $67551.5
7.6
23.4
RTX 4060 Ti 8GB8 GB
about $40029.4
7.4
18.4
RTX 4070 Ti12 GB
about $72251.5
7.1
18.1

Cost per token: the GTX 1660 against 6–12 GB cards

On Qwen3.5 9B (9B params, Q4_K_M, 4K context). Speed, cost per token and power per token depend on the model.

6 GB12 GBthis cardat MSRP
$0.00$0.50$1.00$1.50$0$1,000$ per 1M output tokens · lower is betterused price →RTX 3060 12GB · 12 GB · ≈$329 · $0.76 per 1M tokensRTX 4060 Ti 8GB · 8 GB · ≈$400 · $1.09 per 1M tokensRTX 5060 · 8 GB · ≈$430 · $0.73 per 1M tokensRTX 5070 · 12 GB · ≈$736 · $0.83 per 1M tokensRTX 4070 SUPER · 12 GB · ≈$675 · $1.01 per 1M tokensRTX 4070 Ti · 12 GB · ≈$722 · $1.12 per 1M tokensRTX 4060 · 8 GB · ≈$335 · $0.94 per 1M tokensRTX 4070 · 12 GB · ≈$575 · $0.87 per 1M tokensGTX 1660 · 6 GB · ≈$105 · $0.52 per 1M tokensGTX 1660RTX 5060RTX 5070RTX 4070 SUPERRTX 4070 TiRTX 4070RTX 3060 12GB
$0.00$0.50$1.00$1.50$0$1k$ per 1M output tokens · lower is betterused price →RTX 3060 12GB · 12 GB · ≈$329 · $0.76 per 1M tokensRTX 4060 Ti 8GB · 8 GB · ≈$400 · $1.09 per 1M tokensRTX 5060 · 8 GB · ≈$430 · $0.73 per 1M tokensRTX 5070 · 12 GB · ≈$736 · $0.83 per 1M tokensRTX 4070 SUPER · 12 GB · ≈$675 · $1.01 per 1M tokensRTX 4070 Ti · 12 GB · ≈$722 · $1.12 per 1M tokensRTX 4060 · 8 GB · ≈$335 · $0.94 per 1M tokensRTX 4070 · 12 GB · ≈$575 · $0.87 per 1M tokensGTX 1660 · 6 GB · ≈$105 · $0.52 per 1M tokensGTX 1660RTX 5060RTX 5070
Cost per 1M output tokenson Qwen3.5 9B · card over 3 years + power
$0.52
Card, used price
$0.29
Power
$0.22
3-year cost of ownership$184
Or rent an RTX 3080 Ti in the cloud$0.12/hr

Renting costs as much as buying after ≈900 hours (≈0.6 years at 4 h a day).

Include card price

Each dot is a card; hover or tap it for its numbers, click to open it. Cards without a price for the chosen basis are left out. How we calculate cost per token

Similar GPUs

Cards people weigh against the GTX 1660, each with the one reason it is worth a look. Speeds on Qwen3.5 9B.

2 GB more and 30% fasterRTX 4060 Ti 8GBNVIDIA · 8 GB · Ada Lovelace
▲ 30% speed▲ 2 GBabout $399rent $0.15/hr
6 GB more and 63% fasterRTX 3060 12GBNVIDIA · 12 GB · Ampere
▲ 63% speed▲ 6 GBabout $329rent $0.06/hr
2 GB more and nearly 2× the speedRTX 5060NVIDIA · 8 GB · Blackwell
▲ 103% speed▲ 2 GBabout $460
More efficient per wattRTX 4060NVIDIA · 8 GB · Ada Lovelace
▲ 23% speed▲ 2 GBabout $520rent $0.08/hr
2 GB more and 78% fasterA750Intel · 8 GB · Alchemist
▲ 78% speed▲ 2 GBabout $289
2 GB more and 78% fasterA580Intel · 8 GB · Alchemist
▲ 78% speed▲ 2 GBabout $179

GTX 1660 tokens per second by model size

Output speed depends on memory bandwidth (192 GB/s). Grey rows need more than 6 GB.

  1. 9BQwen3.5 9B · Q3_K_L · 5.8 GB tight · Q3_K_L
    22.6tok/s
  2. 14BQwen3 14B · not at 4-bit
    needs 11.9 GB
  3. 27–31BQwen3.8 27B · not at 4-bit
    needs 17.9 GB
  4. 35B MoEQwen3.6 35B A3B · not at 4-bit
    needs 22.0 GB
  5. 70BLlama 3.3 70B Instruct · not at 4-bit
    needs 51.3 GB
  6. 120B MoEGPT OSS 120B · not at 4-bit
    needs 70.8 GB

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

All 738 models on the GTX 1660

Featured only

Best quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with NVIDIA GeForce GTX 1660 — ranked by Score, recency and Fit
ModelVRAMScoreFit
Mamba 2.8B HF
Q4_K_M·68.2 t/s
3.7 GBS85, Runs great
TinyLlama 1.1B Intermediate Step 1431k 3T
Q4_K_M·123.6 t/s
1.0 GBS95, Runs great
Pythia 1B
Q4_K_M·175.8 t/s
0.7 GBS99, Runs great
Llama2 Kv Sharing
Q4_K_M·416.0 t/s
0.3 GBS100, Runs great
Xglm 564M
Q4_K_M·337.3 t/s
0.4 GBS100, Runs great
DialoGPT Small
Q4_K_M·1040.0 t/s
0.1 GBS100, Runs great
Tiny Qwen3ForCausalLM
Q4_K_M·416.0 t/s
0.6 GBS100, Runs great
Openai GPT
Q4_K_M·1560.0 t/s
0.1 GBS100, Runs great
Pythia 1.4B
Q4_K_M·124.8 t/s
1 GBS95, Runs great
Llama 160M
Q4_K_M·265.5 t/s
0.5 GBS100, Runs great
SmolLM2 1.7B Instruct
Q4_K_M·72.1 t/s
2.9 GBS86, Runs great
Qehwa Pashto Llm
Q4_K_M·25.0 t/s
5.8 GBC39, Usable
Nidum Gemma 2B Uncensored
Q4_K_M·67.8 t/s
1.9 GBS85, Runs great
GLM 4.6V Flash
Q3_K_M·23.1 t/s
6.0 GBC32, Usable
Pythia 410M Deduped
Q4_K_M·378.2 t/s
0.3 GBS100, Runs great

GTX 1660 specs, power and PSU

Specs

Architecture
Turing
VRAM
6 GB GDDR5
Memory bandwidth
192.0 GB/s
FP16 (dense)
10.05 TFLOPS
CUDA cores
1,408
Tensor cores
0
Compute capability
7.5
TDP
120 W
Released
2019-03-14
MSRP
$219

GTX 1660 FAQ

Is the GTX 1660 still worth it in 2026?

At ≈$105 used you get 21.5 tok/s per $100 on Qwen3.5 9B, which ranks 1st of the 8 cards compared here; it is also the 4th fastest out of 6 cards with 6 GB. It runs only small models well; 7–9B only just fits.

What is the largest model the GTX 1660 can run?

Qwen2.5 7B Instruct (7.6B) is the largest at 4-bit or better: at Q4_K_M it uses 5.8 GB of 6 GB, and runs at about 25.0 tok/s. At 3-bit, a bigger model fits too: Gemma 2 9B IT (9.2B) at IQ3_M, 5.7 GB.

Can the GTX 1660 run Qwen3.5 4B?

Yes. Qwen3.5 4B fits at Q4_K_M, using 3.4 GB of 6 GB, and it is the best model that runs well on this card.

Can the GTX 1660 run 14B models?

No. Qwen3 14B needs 11.9 GB at Q4, more than the 6 GB on this card.

How many tokens per second does the GTX 1660 give?

About 38.3 tok/s on Qwen3.5 4B at Q4_K_M, with the first token after 8.4 s on a 4K-token prompt.

What power supply does the GTX 1660 need?

Plan a 250 W PSU (2× the 120 W TDP).

Is renting cheaper than buying a GTX 1660?

An RTX 3080 Ti (12 GB) rents from $0.12/hr. Renting costs as much as buying after ≈900 hours (≈0.6 years at 4 h a day).

Related

Best coding LLMs8 GB VRAMGPUs with 6 GB or moreRTX 3080 TiRTX 3080 12GBRTX 3080