Hardware / Apple /iPhone 17 Pro

Apple · Phone

iPhone 17 Pro

Runs 65 models fully in memory, 0 more with CPU offload. 3 of those sit close enough to 6.6 GB that our size estimate is what decides them.

Memory one app may use

6.6 GB of 12 GB

the most one app may use, not the memory in the phone — estimated

Bandwidth

120 GB/s

Runs

65

01

Specifications

Memory one app may use

6.6 GB of 12 GB

Bandwidth

120 GB/s

FP16 compute

—

TDP

—

MSRP

—

Street price

—

Released

Sep 19, 2025

Architecture

A19 Pro

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Niagara 19m Batch.en20MQ4_K_M~262K4454 tok/sestFits in memory Niagara 38m Batch.en38MQ5_K_M~262K1998 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ8_0~262K169 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ8_0~262K159 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ4_K_M~262K278 tok/sestFits in memory Canary 180m Flash0.2BQ5_K_M~262K422 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ4_K_M~262K139 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ4_K_M~262K149 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ4_K_M~262K149 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ4_K_M~262K149 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ5_K_M~262K690 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ5_K_M~262K127 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ5_K_M~262K127 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ5_K_M~262K262 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ4_K_M~262K307 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ4_K_M~262K186 tok/sestFits in memory Moonshine Streaming Medium0.2BQ5_K_M~262K316 tok/sestFits in memory Moonshine Streaming Small0.1BQ4_K_M~131K891 tok/sestFits in memory Moonshine Streaming Tiny30MQ5_K_M~262K2531 tok/sestFits in memory Moonshine Tiny30MQ4_K_M~262K2969 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ4_K_M~262K114 tok/sestFits in memory Distil Large v3.50.8BQ4_K_M~262K111 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ4_K_M~262K111 tok/sestFits in memory Qwen3 ASR 0.6B HF0.8BQ4_K_M~33K111 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ4_K_M~131K89 tok/sestFits in memory Owsm CTC v4 1B1BQ4_K_M~131K89 tok/sestFits in memory MMS 1b All1BQ4_K_M~131K89 tok/sestFits in memory Canary 1b1BQ4_K_M~131K89 tok/sestFits in memory Canary 1b Flash1BQ4_K_M~131K89 tok/sestFits in memory Canary 1b v21BQ4_K_M~131K89 tok/sestFits in memory MOSS Transcribe Diarize0.9BQ4_K_M~33K99 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ4_K_M~131K81 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ4_K_M~131K81 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ4_K_M~131K81 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ4_K_M~131K81 tok/sestFits in memory ARK ASR 0.6B1.3BQ4_K_M~262K69 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ4_K_M33K74 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ4_K_M~131K64 tok/sestFits in memory Whisper Large v31.5BQ4_K_M~131K59 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ4_K_M~131K56 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ4_K_M~131K42 tok/sestFits in memory Qwen3 ASR 1.7B2BQ4_K_M~131K44 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ4_K_M~131K44 tok/sestFits in memory GLM ASR Nano 25122.3BQ4_K_M~66K39 tok/sestFits in memory STT 2.6b en2.6BQ4_K_M~66K34 tok/sestFits in memory Canary Qwen 2.5b2.6BQ4_K_M~66K34 tok/sestFits in memory Llama 3.2 3B Instruct3.2BQ4_K_M66K28 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ4_K_M~33K39 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ4_K_M~33K39 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ4_K_M~33K39 tok/sestFits in memory Gemma 3 4B4.3BQ4_K_M66K21 tok/sestFits in memory ARK ASR 3B4.1BQ4_K_M~66K22 tok/sestFits in memory Granite Speech 3.3 2b3BQ4_K_M~33K30 tok/sestFits in memory Granite 4.0 Micro3.2BQ4_K_M33K24 tok/sestFits in memory Higgs Audio v3 STT2.7BQ4_K_M~16K33 tok/sestFits in memory Hojo ASR V15.2BQ5_K_M~33K15 tok/sestFits in memory Voxtral Mini 3B 25075BQ4_K_M~33K18 tok/sestFits in memory Ministral 3 3B 25123.8BQ4_K_M16K23 tok/sestFits in memory MOSS Transcribe Preview 2B2.4BQ4_K_M~8K37 tok/sestFits in memory Voxtral Mini 4B Realtime 26024.4BQ4_K_M~16K20 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