Hardware / Apple /Apple M4 Pro (20-core GPU)

Apple · Apple Silicon

Apple M4 Pro (20-core GPU)

for example MacBook Pro 14-inch and 16-inch (2024), Mac mini (2024)

Runs 123 models fully in memory, 10 more with CPU offload. 7 of those sit close enough to 64 GB that our size estimate is what decides them.

Memory

64 GB

Bandwidth

273 GB/s

Runs

123

01

Specifications

Memory

64 GB

Bandwidth

273 GB/s

FP16 compute

17 TFLOPS

TDP

40 W

MSRP

—

Street price

—

Released

Oct 30, 2024

Architecture

M4

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Niagara 19m Batch.en20MQ5_K_M~262K13434 tok/sestFits in memory Niagara 38m Batch.en38MQ4_K_M~262K8295 tok/sestFits in memory ARK ASR 0.6B1.3BQ5_K_M~262K207 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ8_0~262K600 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ4_K_M~262K150 tok/sestFits in memory Distil Large v3.50.8BQ4_K_M~262K394 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ4_K_M~262K225 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ5_K_M~262K244 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ4_K_M~262K315 tok/sestFits in memory Owsm CTC v4 1B1BQ8_0~262K180 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ8_0~262K562 tok/sestFits in memory MMS 1b All1BQ8_0~262K180 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ5_K_M~262K840 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ4_K_M~262K137 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ4_K_M~262K137 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ4_K_M~262K137 tok/sestFits in memory STT 2.6b en2.6BQ4_K_M~262K121 tok/sestFits in memory Canary 180m Flash0.2BQ4_K_M~262K1751 tok/sestFits in memory Canary 1b1BQ5_K_M~262K269 tok/sestFits in memory Canary 1b Flash1BQ5_K_M~262K269 tok/sestFits in memory Canary 1b v21BQ4_K_M~262K315 tok/sestFits in memory Canary Qwen 2.5b2.6BQ4_K_M~262K121 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ8_0~262K281 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ4_K_M~262K525 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ5_K_M~262K448 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ5_K_M~262K244 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ8_0~262K300 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ8_0~262K164 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ8_0~262K1635 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ5_K_M~262K448 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ5_K_M~262K448 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ5_K_M~262K244 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ5_K_M~262K168 tok/sestFits in memory Whisper Large v31.5BQ5_K_M~262K179 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ5_K_M~262K336 tok/sestFits in memory MOSS Transcribe Diarize0.9BQ8_0~262K200 tok/sestFits in memory MOSS Transcribe Preview 2B2.4BQ4_K_M~66K131 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ5_K_M~262K345 tok/sestFits in memory Qwen3 ASR 0.6B HF0.8BQ5_K_M~262K336 tok/sestFits in memory Qwen3 ASR 1.7B2BQ4_K_M~262K155 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ5_K_M~262K132 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ5_K_M~262K927 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ4_K_M~262K1087 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ5_K_M~262K560 tok/sestFits in memory Moonshine Streaming Medium0.2BQ4_K_M~262K1313 tok/sestFits in memory Moonshine Streaming Small0.1BQ5_K_M~262K2687 tok/sestFits in memory Moonshine Streaming Tiny30MQ5_K_M~262K8956 tok/sestFits in memory Moonshine Tiny30MQ8_0~262K5995 tok/sestFits in memory GLM ASR Nano 25122.3BQ4_K_M~262K137 tok/sestFits in memory Higgs Audio v3 STT2.7BQ4_K_M~262K117 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ5_K_M33K224 tok/sestFits in memory Granite Speech 3.3 2b3BQ4_K_M~262K105 tok/sestFits in memory Llama 3.2 3B Instruct3.2BQ4_K_M131K99 tok/sestFits in memory Qwen3.6 35B A3B36BQ4_K_M262K89 tok/sestFits in memory Nemotron 3 Nano 30B A3B31.6BQ4_K_M262K89 tok/sestFits in memory Qwen3 VL 30B A3B Thinking31.1BQ4_K_M262K89 tok/sestFits in memory Qwen3.5-35B-A3B36BQ4_K_M262K89 tok/sestFits in memory Qwen3 VL 30B A3B Instruct31.1BQ4_K_M262K89 tok/sestFits in memory Qwen3 30B A3B Thinking 250730.5BQ4_K_M66K89 tok/sestFits in memory Qwen3 Coder 30B A3B Instruct30.5BQ4_K_M262K89 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