Hardware / NVIDIA /GeForce RTX 3080 10GB

NVIDIA · GPU

GeForce RTX 3080 10GB

Runs 77 models fully in memory, 9 more with CPU offload. 1 of those sit close enough to 10 GB that our size estimate is what decides them.

Memory

10 GB

Bandwidth

760 GB/s

Runs

77

01

Specifications

Memory

10 GB

Bandwidth

760 GB/s

FP16 compute

29.8 TFLOPS

TDP

320 W

MSRP

$699

Street price

—

Released

Sep 17, 2020

Architecture

Ampere

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Granite 4.0 Micro3.2BQ4_K_M66K266 tok/sestFits in memory Llama 3.2 3B Instruct3.2BQ4_K_M131K313 tok/sestFits in memory Gemma 3 4B4.3BQ4_K_M131K233 tok/sestFits in memory Niagara 19m Batch.en20MQ4_K_M~262K50144 tok/sestFits in memory Niagara 38m Batch.en38MQ8_0~262K15059 tok/sestFits in memory ARK ASR 0.6B1.3BQ8_0~262K440 tok/sestFits in memory ARK ASR 3B4.1BQ4_K_M~131K245 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ5_K_M~262K2849 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ8_0~131K273 tok/sestFits in memory Distil Large v3.50.8BQ4_K_M~262K1254 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ5_K_M~262K611 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ8_0~262K520 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ5_K_M~262K855 tok/sestFits in memory Owsm CTC v4 1B1BQ8_0~262K572 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ8_0~262K1788 tok/sestFits in memory MMS 1b All1BQ4_K_M~262K1003 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ8_0~262K1788 tok/sestFits in memory Hojo ASR V15.2BQ4_K_M~66K194 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ8_0~66K249 tok/sestFits in memory Granite Speech 3.3 2b3BQ5_K_M~66K285 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ4_K_M~66K436 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ4_K_M~66K436 tok/sestFits in memory STT 2.6b en2.6BQ5_K_M~131K329 tok/sestFits in memory Voxtral Mini 3B 25075BQ5_K_M~66K171 tok/sestFits in memory Canary 180m Flash0.2BQ4_K_M~262K5572 tok/sestFits in memory Canary 1b1BQ4_K_M~262K1003 tok/sestFits in memory Canary 1b Flash1BQ4_K_M~262K1003 tok/sestFits in memory Canary 1b v21BQ8_0~262K572 tok/sestFits in memory Canary Qwen 2.5b2.6BQ4_K_M~131K386 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ4_K_M~262K1567 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ4_K_M~262K1672 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ8_0~262K954 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ4_K_M~262K912 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ8_0~262K954 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ4_K_M~262K912 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ8_0~262K5202 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ8_0~262K954 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ8_0~262K954 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ8_0~262K520 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ8_0~262K358 tok/sestFits in memory Whisper Large v31.5BQ8_0~262K382 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ5_K_M~262K1069 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ5_K_M~262K1096 tok/sestFits in memory Qwen3 ASR 1.7B2BQ5_K_M~131K419 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ5_K_M~131K419 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ8_0~262K1973 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ5_K_M~262K2948 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ5_K_M~262K1781 tok/sestFits in memory Moonshine Streaming Medium0.2BQ5_K_M~262K3562 tok/sestFits in memory Moonshine Streaming Small0.1BQ8_0~262K5722 tok/sestFits in memory Moonshine Streaming Tiny30MQ4_K_M~262K33430 tok/sestFits in memory Moonshine Tiny30MQ8_0~262K19075 tok/sestFits in memory GLM ASR Nano 25122.3BQ8_0~66K249 tok/sestFits in memory Ministral 3 3B 25123.8BQ8_033K151 tok/sestFits in memory UI-TARS 7B8.3BQ4_K_M33K121 tok/sestFits in memory Gemma 3n 4B7.8BQ4_K_M33K129 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ4_K_M33K836 tok/sestFits in memory Qwen2.5 7B Instruct7.6BQ4_K_M33K132 tok/sestFits in memory Llama 3.1 8B Instruct8BQ4_K_M33K125 tok/sestFits in memory Higgs Audio v3 STT2.7BQ8_0~33K212 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