Hardware / NVIDIA /GeForce GTX 1660 SUPER

NVIDIA · GPU

GeForce GTX 1660 SUPER

Runs 61 models fully in memory, 15 more with CPU offload.

Memory

6 GB

Bandwidth

336 GB/s

Runs

61

01

Specifications

Memory

6 GB

Bandwidth

336 GB/s

FP16 compute

5 TFLOPS

TDP

125 W

MSRP

$229

Street price

—

Released

Oct 29, 2019

Architecture

Turing

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Llama 3.2 3B Instruct3.2BQ4_K_M66K139 tok/sestFits in memory Niagara 19m Batch.en20MQ4_K_M~262K22169 tok/sestFits in memory Niagara 38m Batch.en38MQ8_0~262K6658 tok/sestFits in memory ARK ASR 0.6B1.3BQ5_K_M~262K291 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ8_0~131K843 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ4_K_M~66K211 tok/sestFits in memory Distil Large v3.50.8BQ5_K_M~131K472 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ8_0~66K181 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ4_K_M~131K403 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ4_K_M~131K443 tok/sestFits in memory Owsm CTC v4 1B1BQ8_0~131K253 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ8_0~262K791 tok/sestFits in memory MMS 1b All1BQ8_0~131K253 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ8_0~262K791 tok/sestFits in memory STT 2.6b en2.6BQ5_K_M~66K145 tok/sestFits in memory Canary 180m Flash0.2BQ8_0~262K1406 tok/sestFits in memory Canary 1b1BQ4_K_M~131K443 tok/sestFits in memory Canary 1b Flash1BQ4_K_M~131K443 tok/sestFits in memory Canary 1b v21BQ4_K_M~131K443 tok/sestFits in memory Canary Qwen 2.5b2.6BQ5_K_M~66K145 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ4_K_M~262K693 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ5_K_M~262K630 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ4_K_M~262K739 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ4_K_M~131K403 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ4_K_M~262K739 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ5_K_M~131K344 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ4_K_M~262K4031 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ5_K_M~262K630 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ8_0~131K422 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ8_0~131K230 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ8_0~66K158 tok/sestFits in memory Whisper Large v31.5BQ8_0~66K169 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ5_K_M~131K472 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ5_K_M~131K485 tok/sestFits in memory Qwen3 ASR 1.7B2BQ5_K_M~66K185 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ4_K_M~66K217 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ8_0~262K872 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ4_K_M~262K1529 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ4_K_M~262K924 tok/sestFits in memory Moonshine Streaming Medium0.2BQ5_K_M~262K1575 tok/sestFits in memory Moonshine Streaming Small0.1BQ5_K_M~131K3779 tok/sestFits in memory Moonshine Streaming Tiny30MQ8_0~262K8433 tok/sestFits in memory Moonshine Tiny30MQ5_K_M~262K12598 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ4_K_M33K370 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ4_K_M~33K193 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ4_K_M~33K193 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ4_K_M~33K193 tok/sestFits in memory Qwen3 ASR 0.6B HF0.8BQ4_K_M~33K554 tok/sestFits in memory GLM ASR Nano 25122.3BQ5_K_M~33K164 tok/sestFits in memory ARK ASR 3B4.1BQ4_K_M~33K108 tok/sestFits in memory Higgs Audio v3 STT2.7BQ5_K_M~16K140 tok/sestFits in memory Granite Speech 3.3 2b3BQ5_K_M~16K126 tok/sestFits in memory MOSS Transcribe Diarize0.9BQ8_0~16K281 tok/sestFits in memory Ministral 3 3B 25123.8BQ4_K_M16K117 tok/sestFits in memory Granite 4.0 Micro3.2BQ4_K_M16K118 tok/sestFits in memory Gemma 3 4B4.3BQ4_K_M33K103 tok/sestFits in memory MOSS Transcribe Preview 2B2.4BQ4_K_M~4K185 tok/sestFits in memory Voxtral Mini 3B 25075BQ4_K_M~16K89 tok/sestFits in memory Voxtral Mini 4B Realtime 26024.4BQ4_K_M~8K101 tok/sestFits in memory Hojo ASR V15.2BQ4_K_M~16K86 tok/sestFits in memory

A speed marked est is calculated from this card's memory bandwidth and the size of the weights it has to read per word. It is not a measurement.

Something wrong on this page? Tell us