Hardware / Apple /Apple M4 Max (40-core GPU)

Apple · Apple Silicon

Apple M4 Max (40-core GPU)

for example MacBook Pro 14-inch and 16-inch (2024), Mac Studio (2025). Memory is a build option on these, so not every one of them is sold with 128 GB.

Runs 163 models fully in memory, 2 more with CPU offload. 2 of those sit close enough to 128 GB that our size estimate is what decides them.

Memory

128 GB

Bandwidth

546 GB/s

Runs

163

01

Specifications

Memory

128 GB

Decides which models fit at all.

Memory bandwidth

546 GB/s

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

Compute

34.1 TFLOPS

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

Power draw

70 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

Qwen3.6 35B A3B36B262K152 tok/sFits in memory Gemma 4 26B A4B25.8B262K134 tok/sFits in memory Nemotron 3 Nano 30B A3B31.6B262K152 tok/sFits in memory Ministral 3 3B 25123.8B131K141 tok/sFits in memory gpt-oss-safeguard-20b21.5B131K149 tok/sFits in memory Granite 4.0 Micro3.2B66K197 tok/sFits in memory Qwen3 VL 30B A3B Thinking31.1B262K179 tok/sFits in memory Qwen3 Next 80B A3B Thinking81.3B262K152 tok/sFits in memory Llama 3.2 3B Instruct3.2B131K197 tok/sFits in memory Qwen3.5-35B-A3B36B262K179 tok/sFits in memory Qwen3 VL 30B A3B Instruct31.1B262K152 tok/sFits in memory Qwen3 Next 80B A3B Instruct81.3B262K152 tok/sFits in memory Qwen3 30B A3B Thinking 250730.5B66K152 tok/sFits in memory Qwen3 Coder 30B A3B Instruct30.5B262K179 tok/sFits in memory Qwen3 30B A3B Instruct 250730.5B262K179 tok/sFits in memory Gemma 3 4B4.3B131K147 tok/sFits in memory Nemotron 3.5 Content Safety4.3B131K125 tok/sFits in memory Llama 3.2 1B Instruct1.2B33K300 tok/sFits in memory gpt-oss-120b120B131K105 tok/sFits in memory Hy-MT2-1.8B2B8K268 tok/sFits in memory Hy-MT2-30B-A3B30.1B8K152 tok/sFits in memory Llama 3.1 8B Instruct8B131K79 tok/sFits in memory UI-TARS 7B8.3B66K76 tok/sFits in memory Qwen3 8B8.2B131K77 tok/sFits in memory Granite 4.1 8B8.8B131K72 tok/sFits in memory Ministral 3 8B 25128.9B262K71 tok/sFits in memory Qwen3 VL 8B Thinking8.8B131K72 tok/sFits in memory Qwen3 VL 8B Instruct8.8B262K72 tok/sFits in memory Qwen2.5 7B Instruct7.6B33K83 tok/sFits in memory Granite 4.2 8B8.8B131K72 tok/sFits in memory Gemma 3n 4B7.8B33K81 tok/sFits in memory Reka Edge7.1B16K89 tok/sFits in memory Qwen3.5-9B9.7B262K65 tok/sFits in memory Qwen3.5-122B-A10B125B131K54 tok/sFits in memory Llama Guard 4 12B12B262K53 tok/sFits in memory Gemma 3 12B12.2B131K52 tok/sFits in memory Rocinante 12B12.2B66K52 tok/sFits in memory Mistral Nemo12.2B131K52 tok/sFits in memory Ministral 3 14B 251213.9B262K45 tok/sFits in memory GLM 4.5 Air111B66K45 tok/sFits in memory Qwen3 14B14.8B131K43 tok/sFits in memory Llama 3 8B Lunaris8B8K79 tok/sFits in memory Hunyuan A13B Instruct80.4B131K41 tok/sFits in memory UnslopNemo 12B12.2B33K52 tok/sFits in memory Reka Flash 320.9B66K30 tok/sFits in memory Hy-MT2-7B8B8K67 tok/sFits in memory Cydonia 24B V4.123.6B131K27 tok/sFits in memory Uncensored24B66K26 tok/sFits in memory Mistral Small 3.2 24B24B131K26 tok/sFits in memory Mistral Small 3.1 24B24B66K26 tok/sFits in memory Qwen3.5-27B27.8B262K23 tok/sFits in memory Gemma 3 27B27.4B262K23 tok/sFits in memory Qwen3.6 27B27.8B262K23 tok/sFits in memory Qwen3.8 27B27.8B262K23 tok/sFits in memory Ternary Bonsai 2 27B27B262K23 tok/sFits in memory Gemma 4 31B31.3B66K20 tok/sFits in memory Olmo 3 32B Think32.2B66K20 tok/sFits in memory Qwen3 VL 32B Instruct33.4B131K19 tok/sFits in memory Phi 414.7B16K43 tok/sFits in memory Qwen3 32B32.8B131K19 tok/sFits in memory Muse Glimmer 30B29.8B131K18 tok/sFits in memory Nex-N2-Mini35.1B262K15 tok/sFits in memory GLM 4.7 Flash31.2B131K17 tok/sFits in memory Laguna XS 2.133.4B262K16 tok/sFits in memory Mistral Small 323.6B33K27 tok/sFits in memory Nemotron 3.5 Lightning31.6B262K17 tok/sFits in memory Nex-N2.5-Mini35.1B262K15 tok/sFits in memory MythoMax 13B13B8K49 tok/sFits in memory Qwen3 Coder Next79.7B262K7 tok/sFits in memory Hermes 4 70B70.6B131K9 tok/sFits in memory Qwen2.5 VL 72B Instruct73.4B66K7 tok/sFits in memory Llama 3.3 Euryale 70B70.6B131K9 tok/sFits in memory Qwen2.5 Coder 32B Instruct32.8B33K19 tok/sFits in memory Llama 3.1 Euryale 70B v2.270.6B131K9 tok/sFits in memory Hermes 3 70B Instruct70.6B66K8 tok/sFits in memory Llama 3.1 70B Instruct70.6B131K9 tok/sFits in memory Llama 3.3 70B Instruct70.6B131K9 tok/sFits in memory Nemotron 3 Super124B131K4 tok/sFits in memory GLM 4.6V108B66K4 tok/sFits in memory Command A111B66K6 tok/sFits in memory Skyfall 36B V236.9B33K17 tok/sFits in memory GLM 4.5V108B66K5 tok/sFits in memory Llama 4 Scout109B66K6 tok/sFits in memory Voxtral Small 24B 250724.3B16K22 tok/sFits in memory ReMM SLERP 13B13B4K49 tok/sFits in memory Qwen2.5 72B Instruct72.7B33K9 tok/sFits in memory Mistral Medium 3.5128B33K5 tok/sFits in memory Ling-2.6-flash107B33K5 tok/sFits in memory Devstral 2 2512125B33K5 tok/sFits in memory Gemma 2 27B27.2B8K23 tok/sFits in memory Magnum v4 72B72.7B16K9 tok/sFits in memory Laguna S 2.1118B16K4 tok/sFits in memoryest Mistral Small 4119B16K5 tok/sFits in memory Ling-3.0-flash128B16K4 tok/sFits in memory Mixtral 8x22B Instruct141B8K14 tok/sFits in memoryest R1 Distill Llama 70B70.6B8K9 tok/sFits in memory

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.en20M17986 tok/sFits in memory Niagara 38m Batch.en38M9466 tok/sFits in memory Higgs Audio v3 STT2.7B133 tok/sFits in memory Cohere Transcribe 03 20262.1B171 tok/sFits in memory Distil Large v3.50.8B450 tok/sFits in memory Lite Whisper Large v3 Acc1.4B384 tok/sFits in memory Owsm CTC v3.1 1B1.1B573 tok/sFits in memory Owsm CTC v3.2 ft 1B1B360 tok/sFits in memory Owsm CTC v4 1B1B630 tok/sFits in memory Hubert Large Ls960 ft0.3B1124 tok/sFits in memory MMS 1b All1B537 tok/sFits in memory Wav2vec2 Large 960h Lv60 Self0.3B1124 tok/sFits in memory Hojo ASR V15.2B122 tok/sFits in memory Granite 4.0 1b Speech2.3B156 tok/sFits in memory Granite Speech 3.3 2b3B210 tok/sFits in memory Granite Speech 4.1 2b2.3B274 tok/sFits in memory Granite Speech 4.1 2b NAR2.3B274 tok/sFits in memory STT 2.6b en2.6B243 tok/sFits in memory Voxtral Mini 3B 25074.7B134 tok/sFits in memory Voxtral Mini 4B Realtime 26024.4B122 tok/sFits in memory Canary 180m Flash0.2B2985 tok/sFits in memory Canary 1b1B360 tok/sFits in memory Canary 1b Flash0.8B450 tok/sFits in memory Canary 1b v21B360 tok/sFits in memory Canary Qwen 2.5b2.6B138 tok/sFits in memory Nemotron 3.5 ASR Streaming 0.6b0.6B600 tok/sFits in memory Nemotron Speech Streaming en 0.6b0.6B600 tok/sFits in memory Parakeet CTC 0.6b0.6B1051 tok/sFits in memory Parakeet CTC 1.1b1.1B327 tok/sFits in memory Parakeet RNNT 0.6b0.6B600 tok/sFits in memory Parakeet RNNT 1.1b1.1B573 tok/sFits in memory Parakeet TDT CTC 110m0.1B3270 tok/sFits in memory Parakeet TDT 0.6b v20.6B896 tok/sFits in memory Parakeet TDT 0.6b v30.6B600 tok/sFits in memory Parakeet TDT 1.1b1.1B327 tok/sFits in memory Nyra Health CrisperWhisper1.6B225 tok/sFits in memory Whisper Large v31.5B358 tok/sFits in memory Whisper Large v3 Turbo0.8B788 tok/sFits in memory MOSS Transcribe Diarize0.9B400 tok/sFits in memory Qwen3 ASR 0.6B0.9B701 tok/sFits in memory Qwen3 ASR 0.6B HF0.8B788 tok/sFits in memory Qwen3 ASR 1.7B2.3B234 tok/sFits in memory Qwen3 ASR 1.7B HF2B180 tok/sFits in memory Zipformer cr CTC Transducer XL 290M0.3B1853 tok/sFits in memory Zipformer Transducer XL 290M0.3B1240 tok/sFits in memory ASR Conformer Largescaleasr0.5B1313 tok/sFits in memory GLM ASR Nano 25122.3B274 tok/sFits in memory Granite Speech 5.0 470m Turboctc0.5B1261 tok/sFits in memory Granite Speech 5.0 470m Turboctc nc0.5B1075 tok/sFits in memory Fun ASR Nano 2512 HF0.8B788 tok/sFits in memory Gemma 4 E2B it5.1B124 tok/sFits in memory Niagara 9m Batch.en9M39969 tok/sFits in memory Niagara 84m Batch.en84M7505 tok/sFits in memory Phi 4 Multimodal Instruct5.6B131K113 tok/sFits in memory ARK ASR 0.6B1.3B33K485 tok/sFits in memory ARK ASR 3B4.1B33K154 tok/sFits in memory Audio8 ASR 0.1B0.3B33K1199 tok/sFits in memory MOSS Transcribe Preview 2B2.4B33K263 tok/sFits in memory Gemma 4 E4B it8B131K79 tok/sFits in memory VibeVoice ASR HF8.3B76 tok/sFits in memory Higgs Audio v3 8b STT v28.9B71 tok/sFits in memory Granite Speech 3.3 8b8.6B73 tok/sFits in memory Moonshine Streaming Medium0.3B4K2101 tok/sFits in memory Moonshine Streaming Small0.1B4K3597 tok/sFits in memory Gemma 4 12B it12B53 tok/sFits in memory Moonshine Streaming Tiny0B4Kno speed estimateFits in memory Moonshine Tiny0B194no speed estimateFits in memory

Runs only by spilling into system memory2 modelsThese load, but part of the weights sits in ordinary system memory, which is far slower than the chip.

Step 3.7 Flash201B—no speed estimateSpills to system RAMest Step 3.5 Flash199B—no speed estimateSpills to system RAMest

Too large for this device59 modelsThe weights do not fit even with part of them offloaded to system memory. Closest calls first.

MiniMax M2.7229B—no speed estimateToo largeest Qwen3 VL 235B A22B Instruct236B—no speed estimateToo largeest MiniMax M2.5229B—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 Laguna M.1226B—no speed estimateToo large Nex-N2-Pro397B—no speed estimateToo large Nex-N2.5-Pro397B—no speed estimateToo large Trinity Large Thinking399B—no speed estimateToo large MiniMax M2.1229B—no speed estimateToo large MiniMax M2229B—no speed estimateToo large Hermes 4 405B406B—no speed estimateToo large Hermes 3 405B Instruct406B—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 M3427B—no speed estimateToo large Inkling Small266B—no speed estimateToo large Llama 4 Maverick402B—no speed estimateToo large Qwen3 Coder 480B A35B480B—no speed estimateToo large ERNIE 4.5 VL 424B A47B424B—no speed estimateToo large Hy3299B—no speed estimateToo large Hy3 preview299B—no speed estimateToo large DeepSeek V4 Flash Vision Exp305B—no speed estimateToo large MiniMax-01456B—no speed estimateToo large MiMo-V2.5311B—no speed estimateToo large GLM 4.5358B—no speed estimateToo large Nemotron 3 Ultra561B—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 DeepSeek V3.1 Terminus685B—no speed estimateToo large DeepSeek V3685B—no speed estimateToo large DeepSeek V3.1685B—no speed estimateToo large GLM 5754B—no speed estimateToo large Kimi K2 07111T—no speed estimateToo large Ling-2.6-1T1T—no speed estimateToo large Inkling952B—no speed estimateToo large R1 0528685B—no speed estimateToo large DeepSeek V3 0324685B—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 K2.7 Code1.1T—no speed estimateToo large Kimi K2.61.1T—no speed estimateToo large DeepSeek V4.1 Flash763B—no speed estimateToo large Hy4 preview780B—no speed estimateToo large MiMo-V2.5-Pro1T—no speed estimateToo large Kimi K2 Thinking1.1T—no speed estimateToo large Kimi K2.51.1T—no speed estimateToo large DeepSeek V4 Pro1.6T—no speed estimateToo large Qwen3.8 2.4T A95B2.4T—no speed estimateToo large LongCat 2.01.8T—no speed estimateToo large Kimi K32.8T—no speed estimateToo large

224 models have a verdict on this device: 163 fit in memory, 2 spill into system memory and 59 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 · less memory Apple M4 Pro (20-core GPU)64 GB · less memory

Something wrong on this page? Tell us