Hardware / Android /Android phone · 12 GB · 2023 or newer

Android · Phone

Android phone · 12 GB · 2023 or newer

for example Galaxy S24, Pixel 9 Pro, Xiaomi 14

Runs 61 models fully in memory, 0 more with CPU offload.

Memory one app may use

6 GB of 12 GB

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

Bandwidth

68 GB/s

Runs

61

01

Specifications

Memory one app may use

6 GB of 12 GB

Bandwidth

68 GB/s

FP16 compute

—

TDP

—

MSRP

—

Street price

—

Released

—

Architecture

—

02

What it runs

Best quant per model — sorted by fit score.

ModelParamsBest quantContextSpeedVerdict

Niagara 19m Batch.en20MQ4_K_M~262K2524 tok/sestFits in memory Niagara 38m Batch.en38MQ4_K_M~262K1328 tok/sestFits in memory Audio8 ASR 0.1B0.3BQ4_K_M~262K168 tok/sestFits in memory Hubert Large Ls960 ft0.3BQ5_K_M~262K134 tok/sestFits in memory Wav2vec2 Large 960h Lv60 Self0.3BQ5_K_M~262K134 tok/sestFits in memory Canary 180m Flash0.2BQ4_K_M~262K280 tok/sestFits in memory Parakeet TDT CTC 110m0.1BQ4_K_M~262K459 tok/sestFits in memory Zipformer cr CTC Transducer XL 290M0.3BQ5_K_M~262K148 tok/sestFits in memory Zipformer Transducer XL 290M0.3BQ4_K_M~262K174 tok/sestFits in memory Moonshine Streaming Medium0.2BQ5_K_M~262K179 tok/sestFits in memory Moonshine Streaming Small0.1BQ8_0~131K288 tok/sestFits in memory Moonshine Streaming Tiny30MQ5_K_M~262K1434 tok/sestFits in memory Moonshine Tiny30MQ4_K_M~262K1683 tok/sestFits in memory ASR Conformer Largescaleasr0.5BQ4_K_M~262K105 tok/sestFits in memory Nemotron Speech Streaming en 0.6b0.6BQ4_K_M~262K84 tok/sestFits in memory Parakeet CTC 0.6b0.6BQ4_K_M~262K84 tok/sestFits in memory Parakeet RNNT 0.6b0.6BQ4_K_M~262K84 tok/sestFits in memory Parakeet TDT 0.6b v20.6BQ4_K_M~262K84 tok/sestFits in memory Parakeet TDT 0.6b v30.6BQ4_K_M~262K84 tok/sestFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6BQ4_K_M~262K79 tok/sestFits in memory Distil Large v3.50.8BQ4_K_M~131K63 tok/sestFits in memory Whisper Large v3 Turbo0.8BQ4_K_M~131K63 tok/sestFits in memory Qwen3 ASR 0.6B0.8BQ4_K_M~262K65 tok/sestFits in memory Owsm CTC v3.2 ft 1B1BQ4_K_M~131K51 tok/sestFits in memory Owsm CTC v4 1B1BQ4_K_M~131K51 tok/sestFits in memory MMS 1b All1BQ4_K_M~131K51 tok/sestFits in memory Canary 1b1BQ4_K_M~131K51 tok/sestFits in memory Canary 1b Flash1BQ4_K_M~131K51 tok/sestFits in memory Canary 1b v21BQ4_K_M~131K51 tok/sestFits in memory Qwen3 ASR 0.6B HF0.8BQ4_K_M~33K63 tok/sestFits in memory Owsm CTC v3.1 1B1.1BQ4_K_M~131K46 tok/sestFits in memory Parakeet CTC 1.1b1.1BQ4_K_M~131K46 tok/sestFits in memory Parakeet RNNT 1.1b1.1BQ4_K_M~131K46 tok/sestFits in memory Parakeet TDT 1.1b1.1BQ4_K_M~131K46 tok/sestFits in memory MOSS Transcribe Diarize0.9BQ4_K_M~33K56 tok/sestFits in memory ARK ASR 0.6B1.3BQ4_K_M~262K39 tok/sestFits in memory Lite Whisper Large v3 Acc1.4BQ4_K_M~131K36 tok/sestFits in memory Llama 3.2 1B Instruct1.2BQ4_K_M33K42 tok/sestFits in memory Nyra Health CrisperWhisper1.6BQ4_K_M~131K32 tok/sestFits in memory Whisper Large v31.5BQ4_K_M~131K34 tok/sestFits in memory Cohere Transcribe 03 20262.1BQ4_K_M~66K24 tok/sestFits in memory Qwen3 ASR 1.7B2BQ4_K_M~131K25 tok/sestFits in memory Qwen3 ASR 1.7B HF2BQ4_K_M~131K25 tok/sestFits in memory STT 2.6b en2.6BQ4_K_M~66K19 tok/sestFits in memory Canary Qwen 2.5b2.6BQ4_K_M~66K19 tok/sestFits in memory Llama 3.2 3B Instruct3.2BQ4_K_M66K16 tok/sestFits in memory Granite 4.0 1b Speech2.3BQ4_K_M~33K22 tok/sestFits in memory Granite Speech 4.1 2b2.3BQ5_K_M~33K19 tok/sestFits in memory Granite Speech 4.1 2b NAR2.3BQ5_K_M~33K19 tok/sestFits in memory GLM ASR Nano 25122.3BQ4_K_M~33K22 tok/sestFits in memory Granite Speech 3.3 2b3BQ4_K_M~33K17 tok/sestFits in memory Granite 4.0 Micro3.2BQ4_K_M33K13 tok/sestFits in memory Gemma 3 4B4.3BQ4_K_M33K12 tok/sestFits in memory ARK ASR 3B4.1BQ5_K_M~33K11 tok/sestFits in memory Voxtral Mini 3B 25075BQ4_K_M~33K10 tok/sestFits in memory Higgs Audio v3 STT2.7BQ4_K_M~16K19 tok/sestFits in memory Hojo ASR V15.2BQ4_K_M~33K10 tok/sestFits in memory Ministral 3 3B 25123.8BQ5_K_M16K11 tok/sestFits in memory Voxtral Mini 4B Realtime 26024.4BQ4_K_M~16K12 tok/sestFits in memory Phi 4 Multimodal Instruct5.6BQ4_K_M~8K9 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