Hardware / NVIDIA /GeForce RTX 3060 12GB

NVIDIA · GPU

GeForce RTX 3060 12GB

Runs 83 models fully in memory, 5 more with CPU offload. 1 of those sit close enough to 12 GB that our size estimate is what decides them.

Memory

12 GB

Bandwidth

360 GB/s

Runs

83

01

Specifications

Memory

12 GB

Bandwidth

360 GB/s

FP16 compute

12.7 TFLOPS

TDP

170 W

MSRP

$329

Street price

—

Released

Feb 25, 2021

Architecture

Ampere

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Ministral 3 3B 25123.8BQ4_K_M66K125 tok/sestFits in memory Granite 4.0 Micro3.2BQ4_K_M66K126 tok/sestFits in memory Llama 3.2 3B Instruct3.2BQ4_K_M131K149 tok/sestFits in memory Niagara 19m Batch.en20MQ4_K_M~262K23753 tok/sestFits in memory Niagara 38m Batch.en38MQ4_K_M~262K12501 tok/sestFits in memory ARK ASR 0.6B1.3BQ4_K_M~262K365 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ4_K_M~262K1584 tok/sestFits in memory Higgs Audio v3 STT2.7BQ5_K_M~66K150 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ5_K_M~262K193 tok/sestFits in memory Distil Large v3.50.8BQ4_K_M~262K594 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ5_K_M~262K289 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ8_0~262K246 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ4_K_M~262K475 tok/sestFits in memory Owsm CTC v4 1B1BQ5_K_M~262K405 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ8_0~262K847 tok/sestFits in memory MMS 1b All1BQ8_0~262K271 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ8_0~262K847 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ5_K_M~66K176 tok/sestFits in memory Granite Speech 3.3 2b3BQ4_K_M~66K158 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ5_K_M~66K176 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ5_K_M~66K176 tok/sestFits in memory STT 2.6b en2.6BQ5_K_M~262K156 tok/sestFits in memory Canary 180m Flash0.2BQ4_K_M~262K2639 tok/sestFits in memory Canary 1b1BQ5_K_M~262K405 tok/sestFits in memory Canary 1b Flash1BQ4_K_M~262K475 tok/sestFits in memory Canary 1b v21BQ8_0~262K271 tok/sestFits in memory Canary Qwen 2.5b2.6BQ5_K_M~262K156 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ5_K_M~262K633 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ4_K_M~262K792 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ4_K_M~262K792 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ4_K_M~262K432 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ5_K_M~262K675 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ4_K_M~262K432 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ4_K_M~262K4319 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ8_0~262K452 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ5_K_M~262K675 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ5_K_M~262K368 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ4_K_M~262K297 tok/sestFits in memory Whisper Large v31.5BQ5_K_M~262K270 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ8_0~262K339 tok/sestFits in memory MOSS Transcribe Diarize0.9BQ4_K_M~66K528 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ5_K_M~262K519 tok/sestFits in memory Qwen3 ASR 0.6B HF0.8BQ8_0~66K339 tok/sestFits in memory Qwen3 ASR 1.7B2BQ8_0~262K133 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ8_0~262K133 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ8_0~262K935 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ5_K_M~262K1396 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ8_0~262K565 tok/sestFits in memory Moonshine Streaming Medium0.2BQ8_0~262K1129 tok/sestFits in memory Moonshine Streaming Small0.1BQ5_K_M~262K4049 tok/sestFits in memory Moonshine Streaming Tiny30MQ5_K_M~262K13497 tok/sestFits in memory Moonshine Tiny30MQ8_0~262K9035 tok/sestFits in memory GLM ASR Nano 25122.3BQ5_K_M~131K176 tok/sestFits in memory ARK ASR 3B4.1BQ4_K_M~131K116 tok/sestFits in memory Gemma 3 4B4.3BQ4_K_M131K111 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ8_033K226 tok/sestFits in memory Voxtral Mini 4B Realtime 26024.4BQ4_K_M~66K108 tok/sestFits in memory Voxtral Mini 3B 25075BQ4_K_M~131K95 tok/sestFits in memory MOSS Transcribe Preview 2B2.4BQ5_K_M~16K169 tok/sestFits in memory Hojo ASR V15.2BQ4_K_M~131K92 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