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). Memory is a build option on these, so not every one of them is sold with 64 GB.

Runs 141 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

141

01

Specifications

Memory

64 GB

Decides which models fit at all.

Memory bandwidth

273 GB/s

How fast it can read the model. Sets the speed of the reply.

Compute

17 TFLOPS

Trillions of half-precision sums a second. Sets how fast it reads a long prompt, not how fast it replies.

Power draw

40 W

At full load, for sizing a power supply.

Released

Oct 30, 2024

Architecture

M4

02

What it runs

Best quant per model. Text models first, then speech, then what does not fit.

What is quantisation? →

ModelParamsContextSpeedVerdict

Llama 3.2 1B Instruct1.2B33K150 tok/sFits in memory Granite 4.0 Micro3.2B66K99 tok/sFits in memory Llama 3.2 3B Instruct3.2B131K99 tok/sFits in memory Qwen3.6 35B A3B36B262K89 tok/sFits in memory Nemotron 3 Nano 30B A3B31.6B262K89 tok/sFits in memory Qwen3 VL 30B A3B Thinking31.1B262K89 tok/sFits in memory Qwen3.5-35B-A3B36B262K89 tok/sFits in memory Qwen3 VL 30B A3B Instruct31.1B262K89 tok/sFits in memory Qwen3 30B A3B Thinking 250730.5B66K89 tok/sFits in memory Qwen3 Coder 30B A3B Instruct30.5B262K89 tok/sFits in memory Qwen3 30B A3B Instruct 250730.5B262K89 tok/sFits in memory Ministral 3 3B 25123.8B131K83 tok/sFits in memory Hy-MT2-1.8B2B8K134 tok/sFits in memory gpt-oss-safeguard-20b21.5B131K74 tok/sFits in memory Gemma 3 4B4.3B131K73 tok/sFits in memory Nemotron 3.5 Content Safety4.3B131K73 tok/sFits in memory Gemma 4 26B A4B25.8B66K67 tok/sFits in memory Hy-MT2-30B-A3B30.1B8K89 tok/sFits in memory UI-TARS 7B8.3B66K38 tok/sFits in memory Qwen3 8B8.2B131K38 tok/sFits in memory Llama 3.1 8B Instruct8B131K39 tok/sFits in memory Granite 4.1 8B8.8B131K36 tok/sFits in memory Ministral 3 8B 25128.9B262K35 tok/sFits in memory Qwen3 VL 8B Thinking8.8B131K36 tok/sFits in memory Qwen3 VL 8B Instruct8.8B262K36 tok/sFits in memory Granite 4.2 8B8.8B131K36 tok/sFits in memory Qwen3.5-9B9.7B262K33 tok/sFits in memory Gemma 3n 4B7.8B33K40 tok/sFits in memory Llama Guard 4 12B12B262K26 tok/sFits in memory Qwen2.5 7B Instruct7.6B33K42 tok/sFits in memory Ministral 3 14B 251213.9B131K23 tok/sFits in memory Gemma 3 12B12.2B131K26 tok/sFits in memory Rocinante 12B12.2B66K26 tok/sFits in memory Mistral Nemo12.2B131K26 tok/sFits in memory Reka Edge7.1B16K44 tok/sFits in memory Qwen3 14B14.8B131K21 tok/sFits in memory Reka Flash 320.9B66K15 tok/sFits in memory Qwen3.5-27B27.8B66K11 tok/sFits in memory Cydonia 24B V4.123.6B131K13 tok/sFits in memory Uncensored24B66K11 tok/sFits in memory Mistral Small 3.2 24B24B131K11 tok/sFits in memory Gemma 3 27B27.4B262K12 tok/sFits in memory UnslopNemo 12B12.2B33K22 tok/sFits in memory Qwen3.6 27B27.8B66K11 tok/sFits in memory Mistral Small 3.1 24B24B66K13 tok/sFits in memory Muse Glimmer 30B29.8B131K11 tok/sFits in memory Qwen3.8 27B27.8B66K11 tok/sFits in memory Ternary Bonsai 2 27B27B262K10 tok/sFits in memory Nex-N2-Mini35.1B131K7 tok/sFits in memory GLM 4.7 Flash31.2B66K9 tok/sFits in memory Olmo 3 32B Think32.2B66K8 tok/sFits in memory Qwen3 VL 32B Instruct33.4B66K8 tok/sFits in memory Laguna XS 2.133.4B66K7 tok/sFits in memory Qwen3 32B32.8B66K10 tok/sFits in memory Nemotron 3.5 Lightning31.6B262K7 tok/sFits in memory Nex-N2.5-Mini35.1B131K7 tok/sFits in memory Llama 3 8B Lunaris8B8K39 tok/sFits in memory Mistral Small 323.6B33K11 tok/sFits in memory Skyfall 36B V236.9B33K9 tok/sFits in memory Phi 414.7B16K18 tok/sFits in memory Qwen2.5 Coder 32B Instruct32.8B33K10 tok/sFits in memory Hy-MT2-7B8B8K34 tok/sFits in memory Gemma 4 31B31.3B16K10 tok/sFits in memory Voxtral Small 24B 250724.3B16K11 tok/sFits in memory MythoMax 13B13B8K24 tok/sFits in memory ReMM SLERP 13B13B4K24 tok/sFits in memory Gemma 2 27B27.2B8K12 tok/sFits in memory Llama 3.1 70B Instruct70.6B4K5 tok/sFits in memoryest Llama 3.3 70B Instruct70.6B4K5 tok/sFits in memoryest Hermes 4 70B70.6B2K5 tok/sFits in memoryest Llama 3.3 Euryale 70B70.6B2K5 tok/sFits in memoryest Llama 3.1 Euryale 70B v2.270.6B2K5 tok/sFits in memoryest R1 Distill Llama 70B70.6B2K5 tok/sFits in memoryest Hermes 3 70B Instruct70.6B2K5 tok/sFits in memoryest

Speech models it runs67 modelsTranscription and speech models that fit in memory. Throughput figures for a transcription model come from its text decoder, so treat this as a fit answer rather than a speed one.

Niagara 19m Batch.en20M13434 tok/sFits in memory Niagara 38m Batch.en38M7071 tok/sFits in memory Cohere Transcribe 03 20262.1B128 tok/sFits in memory Distil Large v3.50.8B394 tok/sFits in memory Lite Whisper Large v3 Acc1.4B225 tok/sFits in memory Owsm CTC v3.1 1B1.1B244 tok/sFits in memory Owsm CTC v3.2 ft 1B1B315 tok/sFits in memory Owsm CTC v4 1B1B269 tok/sFits in memory Hubert Large Ls960 ft0.3B840 tok/sFits in memory MMS 1b All1B180 tok/sFits in memory Wav2vec2 Large 960h Lv60 Self0.3B562 tok/sFits in memory Granite 4.0 1b Speech2.3B137 tok/sFits in memory Granite Speech 4.1 2b2.3B137 tok/sFits in memory Granite Speech 4.1 2b NAR2.3B137 tok/sFits in memory STT 2.6b en2.6B121 tok/sFits in memory Canary 180m Flash0.2B1751 tok/sFits in memory Canary 1b1B180 tok/sFits in memory Canary 1b Flash0.8B225 tok/sFits in memory Canary 1b v21B180 tok/sFits in memory Canary Qwen 2.5b2.6B121 tok/sFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6B300 tok/sFits in memory Nemotron Speech Streaming en 0.6b0.6B300 tok/sFits in memory Parakeet CTC 0.6b0.6B525 tok/sFits in memory Parakeet CTC 1.1b1.1B164 tok/sFits in memory Parakeet RNNT 0.6b0.6B300 tok/sFits in memory Parakeet RNNT 1.1b1.1B287 tok/sFits in memory Parakeet TDT CTC 110m0.1B1635 tok/sFits in memory Parakeet TDT 0.6b v20.6B448 tok/sFits in memory Parakeet TDT 0.6b v30.6B525 tok/sFits in memory Parakeet TDT 1.1b1.1B164 tok/sFits in memory Nyra Health CrisperWhisper1.6B197 tok/sFits in memory Whisper Large v31.5B120 tok/sFits in memory Whisper Large v3 Turbo0.8B336 tok/sFits in memory MOSS Transcribe Diarize0.9B200 tok/sFits in memory Qwen3 ASR 0.6B0.9B350 tok/sFits in memory Qwen3 ASR 0.6B HF0.8B394 tok/sFits in memory Qwen3 ASR 1.7B2.3B137 tok/sFits in memory Qwen3 ASR 1.7B HF2B158 tok/sFits in memory Zipformer cr CTC Transducer XL 290M0.3B927 tok/sFits in memory Zipformer Transducer XL 290M0.3B620 tok/sFits in memory ASR Conformer Largescaleasr0.5B657 tok/sFits in memory GLM ASR Nano 25122.3B137 tok/sFits in memory Granite Speech 5.0 470m Turboctc0.5B537 tok/sFits in memory Granite Speech 5.0 470m Turboctc nc0.5B360 tok/sFits in memory Fun ASR Nano 2512 HF0.8B394 tok/sFits in memory Niagara 9m Batch.en9M19984 tok/sFits in memory Niagara 84m Batch.en84M2141 tok/sFits in memory Higgs Audio v3 STT2.7B117 tok/sFits in memory ARK ASR 0.6B1.3B33K243 tok/sFits in memory Audio8 ASR 0.1B0.3B33K896 tok/sFits in memory MOSS Transcribe Preview 2B2.4B33K131 tok/sFits in memory Granite Speech 3.3 2b3B105 tok/sFits in memory Voxtral Mini 4B Realtime 26024.4B72 tok/sFits in memory Moonshine Streaming Medium0.3B4K1051 tok/sFits in memory Moonshine Streaming Small0.1B4K1799 tok/sFits in memory Voxtral Mini 3B 25074.7B67 tok/sFits in memory ARK ASR 3B4.1B33K77 tok/sFits in memory Hojo ASR V15.2B61 tok/sFits in memory Gemma 4 E2B it5.1B62 tok/sFits in memory Phi 4 Multimodal Instruct5.6B131K56 tok/sFits in memory VibeVoice ASR HF8.3B38 tok/sFits in memory Gemma 4 E4B it8B131K39 tok/sFits in memory Higgs Audio v3 8b STT v28.9B35 tok/sFits in memory Granite Speech 3.3 8b8.6B37 tok/sFits in memory Gemma 4 12B it12B26 tok/sFits in memory Moonshine Streaming Tiny0B4Kno speed estimateFits in memory Moonshine Tiny0B194no speed estimateFits in memory

Runs only by spilling into system memory10 modelsThese load, but part of the weights sits in ordinary system memory, which is far slower than the chip.

Ling-2.6-flash107B—no speed estimateSpills to system RAMest Qwen3 Coder Next79.7B—no speed estimateSpills to system RAM GLM 4.6V108B—no speed estimateSpills to system RAMest Qwen3 Next 80B A3B Thinking81.3B—no speed estimateSpills to system RAM Hunyuan A13B Instruct80.4B—no speed estimateSpills to system RAM Qwen2.5 VL 72B Instruct73.4B—no speed estimateSpills to system RAM Magnum v4 72B72.7B—no speed estimateSpills to system RAMest Qwen3 Next 80B A3B Instruct81.3B—no speed estimateSpills to system RAM GLM 4.5V108B—no speed estimateSpills to system RAMest Qwen2.5 72B Instruct72.7B—no speed estimateSpills to system RAM

Too large for this device73 modelsThe weights do not fit even with part of them offloaded to system memory. Closest calls first.

Nemotron 3 Super124B—no speed estimateToo large GLM 4.5 Air111B—no speed estimateToo large gpt-oss-120b120B—no speed estimateToo large Mistral Small 4119B—no speed estimateToo large Devstral 2 2512125B—no speed estimateToo large Mistral Medium 3.5128B—no speed estimateToo large Laguna S 2.1118B—no speed estimateToo large Mixtral 8x22B Instruct141B—no speed estimateToo large Llama 4 Scout109B—no speed estimateToo large Command A111B—no speed estimateToo large Qwen3.5-122B-A10B125B—no speed estimateToo large Ling-3.0-flash128B—no speed estimateToo large MiniMax M2.7229B—no speed estimateToo large Step 3.5 Flash199B—no speed estimateToo large Qwen3 VL 235B A22B Instruct236B—no speed estimateToo large Step 3.7 Flash201B—no speed estimateToo large Inkling Small266B—no speed estimateToo large MiniMax M2.5229B—no speed estimateToo large MiniMax M2229B—no speed estimateToo large Qwen3 235B A22B Instruct 2507235B—no speed estimateToo large Qwen3 VL 235B A22B Thinking236B—no speed estimateToo large DeepSeek V4 Flash291B—no speed estimateToo large GLM 5.3 Flash321B—no speed estimateToo large Hy3299B—no speed estimateToo large Laguna M.1226B—no speed estimateToo large MiniMax M2.1229B—no speed estimateToo large Qwen3 235B A22B Thinking 2507235B—no speed estimateToo large GLM 4.6357B—no speed estimateToo large GLM 4.7358B—no speed estimateToo large MiniMax-01456B—no speed estimateToo large Nex-N2.5-Pro397B—no speed estimateToo large Trinity Large Thinking399B—no speed estimateToo large Llama 4 Maverick402B—no speed estimateToo large Hermes 4 405B406B—no speed estimateToo large Hermes 3 405B Instruct406B—no speed estimateToo large Qwen3 Coder 480B A35B480B—no speed estimateToo large ERNIE 4.5 VL 424B A47B424B—no speed estimateToo large MiniMax M3427B—no speed estimateToo large Hy3 preview299B—no speed estimateToo large DeepSeek V4 Flash Vision Exp305B—no speed estimateToo large MiMo-V2.5311B—no speed estimateToo large Nemotron 3 Ultra561B—no speed estimateToo large GLM 4.5358B—no speed estimateToo large Nex-N2-Pro397B—no speed estimateToo large Jamba Large 1.7399B—no speed estimateToo large Qwen3.5 397B A17B403B—no speed estimateToo large GLM 5.1754B—no speed estimateToo large GLM 5.2753B—no speed estimateToo large MiniMax M1456B—no speed estimateToo large R1 0528685B—no speed estimateToo large DeepSeek V3.1 Terminus685B—no speed estimateToo large DeepSeek V3685B—no speed estimateToo large GLM 5754B—no speed estimateToo large DeepSeek V4.1 Flash763B—no speed estimateToo large Inkling952B—no speed estimateToo large Kimi K2 07111T—no speed estimateToo large Ling-2.6-1T1T—no speed estimateToo large DeepSeek V3 0324685B—no speed estimateToo large DeepSeek V3.1685B—no speed estimateToo large DeepSeek V3.2685B—no speed estimateToo large DeepSeek V3.2 Exp685B—no speed estimateToo large Kimi K2 09051T—no speed estimateToo large Ring-2.6-1T1T—no speed estimateToo large Kimi K2.61.1T—no speed estimateToo large Kimi K2.51.1T—no speed estimateToo large Hy4 preview780B—no speed estimateToo large DeepSeek V4 Pro1.6T—no speed estimateToo large LongCat 2.01.8T—no speed estimateToo large MiMo-V2.5-Pro1T—no speed estimateToo large Kimi K2 Thinking1.1T—no speed estimateToo large Kimi K2.7 Code1.1T—no speed estimateToo large Qwen3.8 2.4T A95B2.4T—no speed estimateToo large Kimi K32.8T—no speed estimateToo large

224 models have a verdict on this device: 141 fit in memory, 10 spill into system memory and 73 are too large. Every one of them is on this page.

03

Other M4 chips

The same generation in other memory sizes. Memory is what decides which models fit.

Apple M4 (10-core GPU)32 GB · less memory Apple M4 Max (32-core GPU)64 GB · same memory Apple M4 Max (40-core GPU)128 GB · more memory

Something wrong on this page? Tell us