MacBook Pro 16-inch · M5 Pro · 18-core CPU / 20-core GPU · 24 GB unified
Apple · MacBook Pro 16-inch · 2026
125 text models run well · 307 GB/s bandwidth
Models on this hardware
Start here
File Q4_K_M · 0.8 GB · bartowski/Llama-3.2-1B-Instruct-GGUF · What is Q4?
Memory Needs ~1 GB · ~7 GB on MacBook Pro 16-inch 2026 · M5 Pro · 24 GB unified, including ~6 GB the system keeps
ollama run hf.co/bartowski/Llama-3.2-1B-Instruct-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
File Q4_K_M · 2.2 GB · ibm-granite/granite-4.2-3b-GGUF · What is Q4?
Memory Needs ~3 GB · ~9 GB on MacBook Pro 16-inch 2026 · M5 Pro · 24 GB unified, including ~6 GB the system keeps
ollama run hf.co/ibm-granite/granite-4.2-3b-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
File Q4_K_M · 5.3 GB · ibm-granite/granite-4.2-8b-GGUF · What is Q4?
Memory Needs ~6 GB · ~12 GB on MacBook Pro 16-inch 2026 · M5 Pro · 24 GB unified, including ~6 GB the system keeps
ollama run hf.co/ibm-granite/granite-4.2-8b-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
Text
Text · LMArena
rating · higher is better
- Color = publisher
Source: LMArena 2026-10-22 · collected 2026-10-25 · dataset · CC BY 4.0 · Models without a rating are left out, not scored as zero.
| gpt-oss-20bQ4 · published params · source | 1,317 ± 6 | ~22 GBTight fit: Needs ~22 GB: runs well, about 2 GB left beyond the system reserve | Q4 · published params |
|---|---|---|---|
| Granite 4.2 3BQ4 · published params · source | 1,290 ± 12 | ~9 GB | Q4 · published params |
| Granite 4.2 8BQ4 · published params · source | 1,290 ± 11 | ~12 GB | Q4 · published params |
| Phi-4Q4 · published params · source | 1,256 ± 5 | ~17 GB | Q4 · published params |
| Llama 3.1 8B InstructQ4 · published params · source | 1,211 ± 4 | ~12 GB | Q4 · published params |
| Llama 3.2 3B InstructQ4 · published params · source | 1,167 ± 8 | ~9 GB | Q4 · published params |
| Llama 3.2 1B InstructQ4 · published params · source | 1,111 ± 8 | ~7 GB | Q4 · published params |
| Qwen3 14BQ4 · published params · source | Not on the LMArena text leaderboard of 2026-10-22 | ~18 GB | Q4 · published params |
No model within 24 GB has a LiveBench score, so the table has no LiveBench column. LMArena ratings in this table: LMArena 2026-10-22 · collected 2026-10-25 · dataset · CC BY 4.0
Tight fit: it runs, but leaves less than 2.4 GB free beyond the 6 GB the system keeps (the memory figure already includes that reserve).
Every reviewed text model here 23
Every one that runs here, well or slowly, scored or not; 7 of 23 have a LiveBench or LMArena result. Run well first, then tight fits, then slow; within each, by Hub downloads over 30 days, read 2026-10-25, not by quality.
| Qwen3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~18 GB | 9.0 | Apache 2.0 | 10.4M |
|---|---|---|---|---|---|---|
| Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~13 GB | 5.0 | Apache 2.0 | 8.1M |
| Qwen3.5 9B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~13 GB | 5.7 | Apache 2.0 | 6.1M |
| Llama 3.1 8B Instruct | Not in LiveBench 2026-06-25 | 1,211 | ~12 GB | 4.9 | Llama 3.1 | 4.8M |
| Qwen3 1.7B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~8 GB | 1.1 | Apache 2.0 | 4.1M |
| Gemma 4 E4B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~12 GB | 5.0 | Apache 2.0 | 4M |
| Llama 3.2 1B Instruct | Not in LiveBench 2026-06-25 | 1,111 | ~7 GB | 0.8 | Llama 3.2 | 3.5M |
| Gemma 4 E2B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~10 GB | 3.1 | Apache 2.0 | 3M |
| Qwen3 4B Instruct 2507 | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 2.5 | Apache 2.0 | 2.6M |
| Gemma 4 12B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~15 GB | 7.1 | Apache 2.0 | 1.7M |
| Llama 3.2 3B Instruct | Not in LiveBench 2026-06-25 | 1,167 | ~9 GB | 2.0 | Llama 3.2 | 1.1M |
| SmolLM3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 1.9 | Apache 2.0 | 718.7K |
| Phi-4 | Not in LiveBench 2026-06-25 | 1,256 | ~17 GB | 8.0 | MIT | 368.3K |
| DeepSeek R1 0528 Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~13 GB | 5.0 | MIT | 355.8K |
| Phi-4 mini | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 2.5 | MIT | 342.8K |
| Olmo 3 7B Instruct | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~12 GB | 4.5 | Apache 2.0 | 262.5K |
| Ministral 3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~17 GB | 8.2 | Apache 2.0 | 126.5K |
| Ministral 3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 2.1 | Apache 2.0 | 87.4K |
| Ministral 3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~13 GB | 5.2 | Apache 2.0 | 79.4K |
| Granite 4.2 8B | Not in LiveBench 2026-06-25 | 1,290 | ~12 GB | 5.3 | Apache 2.0 | 78.3K |
| Granite 4.2 3B | Not in LiveBench 2026-06-25 | 1,290 | ~9 GB | 2.2 | Apache 2.0 | 39.8K |
| gpt-oss-20b | Not in LiveBench 2026-06-25 | 1,317 | ~22 GBTight fit | 11.6 | Apache 2.0 | 5.8M |
| Devstral Small 2 24B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~24 GBTight fit | 14.3 | Apache 2.0 | 195.6K |
With 32 GB you also run… 17
The next memory size on this ladder runs these well. Figures are the requirement on this hardware type.
- Gemma 4 26B A4BGoogle~25 GBQ4 · published params
- Mistral Small 3.2 24BMistral~25 GBQ4 · tensor count
- Qwen3.8 27BAlibaba~27 GBQ4 · published params
- Qwen3.6 27BAlibaba~27 GBQ4 · published params
- Smaug MiniAbacus AI~27 GBQ4 · published params
- Qwen3 30B A3B Instruct 2507Alibaba~29 GBQ4 · published params
- GLM-4.7 FlashZ.ai~29 GBQ4 · published params
- Nemotron 3.5 Lightning 30B A3BNVIDIA~29 GBQ4 · published params
- Qwen3 Coder 30B A3BAlibaba~29 GBQ4 · published params
- Granite 4.2 30BIBM~29 GBQ4 · published params
- Gemma 4 31BGoogle~30 GBQ4 · published params
- Olmo 3 32B ThinkAi2~30 GBQ4 · published params
- txgemma-27b-chatgoogle~27 GBQ4 · tensor count
- txgemma-27b-predictgoogle~27 GBQ4 · tensor count
- granite-4.1-30bibm-granite~28 GBQ4 · tensor count
- granite-4.1-30b-baseibm-granite~28 GBQ4 · tensor count
- Qwen3-30B-A3B-Thinking-2507Qwen~29 GBQ4 · tensor count
Adjust the estimate: quantization, context, reserve
Vision
Vision · LMArena
rating · higher is better
- Color = publisher
Source: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0 · Models without a rating are left out, not scored as zero.
| Molmo2-8BQ4 · Hugging Face count · source | 1,112 ± 20 | ~13 GB | Q4 · Hugging Face count |
|---|---|---|---|
| Qwen2-VL-7B-InstructQ4 · Hugging Face count · source | 1,031 ± 11 | ~13 GB | Q4 · Hugging Face count |
| Molmo-7B-D-0924Q4 · Hugging Face count · source | 994 ± 14 | ~12 GB | Q4 · Hugging Face count |
| Llama-3.2-11B-Vision-InstructQ4 · Hugging Face count · source | 993 ± 11 | ~15 GB | Q4 · Hugging Face count |
| MiniCPM-V-2_6Q4 · Hugging Face count · source | 962 ± 16 | ~13 GB | Q4 · Hugging Face count |
| Phi-3.5-vision-instructQ4 · Hugging Face count · source | 919 ± 16 | ~10 GB | Q4 · Hugging Face count |
| Gemma 4 12BQ4 · published params · source | Not on the LMArena vision leaderboard of 2026-10-19 | ~15 GB | Q4 · published params |
| Qwen3.5 9BQ4 · published params · source | Not on the LMArena vision leaderboard of 2026-10-19 | ~13 GB | Q4 · published params |
LMArena ratings in this table: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0
With 32 GB you also run… 13
The next memory size on this ladder runs these well. Figures are the requirement on this hardware type.
- Gemma 4 26B A4BGoogle~25 GBQ4 · published params
- Qwen3.6 27BAlibaba~27 GBQ4 · published params
- Qwen3.8 27BAlibaba~27 GBQ4 · published params
- Smaug MiniAbacus AI~27 GBQ4 · published params
- Gemma 4 31BGoogle~30 GBQ4 · published params
- diffusiongemma-26B-A4B-itgoogle~26 GBQ4 · Hugging Face count
- deepseek-vl2deepseek-ai~27 GBQ4 · Hugging Face count
- Fara1.5-27Bmicrosoft~27 GBQ4 · Hugging Face count
- gemma-3-27b-itgoogle~27 GBQ4 · Hugging Face count
- ERNIE-4.5-VL-28B-A3B-PTbaidu~29 GBQ4 · Hugging Face count
- ERNIE-4.5-VL-28B-A3B-Thinkingbaidu~29 GBQ4 · Hugging Face count
- Keye-VL-2.0-30B-A3BKwai-Keye~30 GBQ4 · Hugging Face count
- Qwen3-VL-30B-A3B-InstructQwen~30 GBQ4 · Hugging Face count
12 more vision models in the catalog have no usable memory figure or parameter count for this profile. Browse Vision models
Image
Image · Prompt alignment · GenEval
0–1 · higher is better
Source: GenEval · 2023 · reviewed 2026-10-05 · Published quality in the original benchmark preset; not measured on this hardware.
| Stable Diffusion XL 1.0Platform path documented · Apple Core ML · source | 0.55 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|---|
| Stable Diffusion 1.5Platform path documented · Apple Core ML · source | 0.43 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
| FLUX.2 [klein] 9B16-bit est. · params · source | No comparable published result | 1,070 ± 3 | ~24 GB | 16-bit est. · params |
| FLUX.2 [klein] Base 9B16-bit est. · params · source | No comparable published result | Arena rated flux-2-klein-9b; this entry is the undistilled Base checkpoint | ~24 GB | 16-bit est. · params |
| FLUX.2-klein-9b-kv16-bit est. · params · source | No comparable published result | Not reviewed for LMArena | ~24 GB | 16-bit est. · params |
| HiDream-O1-Image-Dev-260416-bit est. · params · source | No comparable published result | Not reviewed for LMArena | ~24 GB | 16-bit est. · params |
| ERNIE Image16-bit est. · params · source | No comparable published result | Not on the LMArena text-to-image leaderboard of 2026-10-07 | ~22 GB | 16-bit est. · params |
| ERNIE Image Turbo16-bit est. · params · source | No comparable published result | Not on the LMArena text-to-image leaderboard of 2026-10-07 | ~22 GB | 16-bit est. · params |
LMArena ratings in this table: LMArena 2026-10-07 · collected 2026-10-25 · dataset · CC BY 4.0
With 32 GB you also run… 10
The next memory size on this ladder. Figures are the requirement on this hardware type.
- Ideogram 4.0Ideogram~25 GB16-bit est. · params
- Ideogram 4.0 (Quality)Ideogram~25 GB16-bit est. · params
- Ideogram 4.0 FastFal~25 GB16-bit est. · params
- Ideogram 4.0 Fast (Quality)Fal~25 GB16-bit est. · params
- Ideogram 4.0 InstantFal~25 GB16-bit est. · params
- Qwen-Image-2.1-PE-T2IQwen~25 GB16-bit est. · params
- FLUX.1 Krea [dev]Black Forest Labs~31 GB16-bit est. · params
- FLUX.1-Kontext-devblack-forest-labs~31 GB16-bit est. · params
- SRPOTencent~31 GB16-bit est. · params
- Step1X-Edit-v1p2stepfun-ai~32 GB16-bit est. · params
17 more image models in the catalog have no usable memory figure or parameter count for this profile. Browse Image models
Video
| HunyuanVideo-1.516-bit est. · params · source | ~22 GB | 16-bit est. · params |
|---|---|---|
| CogVideoX-5b-I2V16-bit est. · params · source | ~16 GB | 16-bit est. · params |
| CogVideoX1.5-5B16-bit est. · params · source | ~16 GB | 16-bit est. · params |
| CogVideoX1.5-5B-I2V16-bit est. · params · source | ~16 GB | 16-bit est. · params |
| CogVideoX-5B16-bit est. · params · source | ~14 GB | 16-bit est. · params |
| Wan2.2-TI2V-5B16-bit est. · params · source | ~14 GB | 16-bit est. · params |
| Wan2.2-TI2V-5B-Diffusers16-bit est. · params · source | ~14 GB | 16-bit est. · params |
| SANA-Video_2.0_5B_720p16-bit est. · params · source | ~13 GB | 16-bit est. · params |
How this list works Documented execution paths on this hardware first, then publisher minimums (CUDA or Apple MPS).
With 32 GB you also run… 2
The next memory size on this ladder. Figures are the requirement on this hardware type.
- mochi-1-previewgenmo~26 GB16-bit est. · params
- Open-Sora-v2hpcai-tech~29 GB16-bit est. · params
21 more video models in the catalog have no usable memory figure or parameter count for this profile. Browse Video models
Audio
Audio · Word error rate · Open ASR
% · lower is better
Source: Word error rate · Open ASR · reviewed 2026-10-20 · Published quality in the original benchmark preset; not measured on this hardware.
| Whisper large-v3-turboPlatform path documented · whisper.cpp / Metal · source | 7.83 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|
| Whisper large-v3Platform path documented · whisper.cpp / Metal · source | 7.44 | Documented path without a memory figure | Platform path documented |
| Llasa-8B16-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
| MOSS-TTS-v1.516-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
| personaplex-7b-v116-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
| Qwen2-Audio-7B-Instruct16-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
| VibeVoice-ASR16-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
| VibeVoice-ASR-Streaming-7B16-bit est. · params · source | No comparable published result | ~23 GB | 16-bit est. · params |
With 32 GB you also run… 2
The next memory size on this ladder. Figures are the requirement on this hardware type.
- Kimi-Audio-7Bmoonshotai~26 GB16-bit est. · params
- Kimi-Audio-7B-Instructmoonshotai~26 GB16-bit est. · params
35 more audio models in the catalog have no usable memory figure or parameter count for this profile. Browse Audio models
Embeddings
Embeddings · MTEB (Multilingual, v2) · mean(task)
0–100 · higher is better
- Color = publisher
| Qwen3-Embedding 8B8-bit · published params · source | 70.6 | ~12 GB | 8-bit · published params |
|---|---|---|---|
| Qwen3-Embedding 4B8-bit · published params · source | 69.5 | ~7 GB | 8-bit · published params |
| Harrier OSS v1 (0.6B)8-bit · published params · source | 69.0 | ~3 GB | 8-bit · published params |
| Qwen3-Embedding 0.6B8-bit · published params · source | 64.3 | ~3 GB | 8-bit · published params |
| Multilingual E5 Large Instruct8-bit · published params · source | 63.2 | ~3 GB | 8-bit · published params |
| EmbeddingGemma 28-bit · published params · source | 61.4 | ~3 GB | 8-bit · published params |
| EmbeddingGemma (300M)8-bit · published params · source | 61.1 | ~3 GB | 8-bit · published params |
| BGE-M38-bit · published params · source | 59.6 | ~3 GB | 8-bit · published params |
Runs well within 24 GB of unified memory · planning estimates, not measured speed
Measurements
Published tokens per second, exactly as measured; never estimated.
| Llama 7B Q4_0 (llama-bench reference model) | 66.33 | Q4_0 | Metal | llama.cpp Apple Silicon summary · community result |
|---|
Specifications
38 sourced fields; the source is named once per group, and a field from another source links to it. Official specifications
Marker: the lightest text model that does not run well here but runs well with 32 GB. No text model that runs well here has a LiveBench score. Q4 planning estimate, not a measured fit.
Reviewed · Manufacturer source
- CPU cores
- 18
- Processor
- Apple M5 Pro
- GPU cores
- 20
- Graphics processor
- Apple M5 Pro GPU
- GPU runtime status
- Not stated by the source
- Memory capacity
- 24 GB
- Memory kind
- Unified memory
- Memory bandwidth
- 307 GB/s
- Battery capacity
- 100 Wh
- Bluetooth
- Bluetooth 6
- Camera features
- Center Stage; Desk View
- Camera resolution
- 12 MP
- Camera video resolution
- 1080p HD
- CPU performance cores
- 12
- CPU super cores
- 6
- Dimensions (cm)
- 35.57 × 24.81 × 1.68 (W × D × H)
- Display color
- 1 billion colors · P3 wide color · True Tone
- Display contrast ratio
- 1,000,000:1
- Display size
- 16.2 in
- Peak display brightness
- 1,600 nits
- Display refresh rate
- 120 Hz
- Display resolution
- 3456 × 2234
- Sustained display brightness
- 1,000 nits
- Maximum external displays
- 3
- Included power adapter
- 140 W
- Neural Engine cores
- 16
- Power frequency
- 50–60 Hz
- Operating system
- macOS
- Supply voltage
- 100–240 V AC
- Ports
- 3 × Thunderbolt 5 (USB-C) · HDMI · SDXC · MagSafe 3 · 3.5 mm headphone jack
- Speaker system
- High-fidelity six-speaker system with force-cancelling woofers
- Storage options
- Base storage: 1 TB; configure-to-order options: 2 TB or 4 TB.
- Maximum Thunderbolt rate
- 120 Gb/s
- Maximum USB4 rate
- 120 Gb/s
- Weight
- 2.14 kg
- Wi-Fi
- Wi-Fi 7 (802.11be)
- Introduced
- 2026
Published 2026 MacBook Pro 16-inch M5 Pro base memory tier. Unified memory is shared with macOS and apps; Apple specifications do not establish local AI runtime support or speed.
Photo: Apple family photo · memory configuration is not pictured. · Manufacturer source
Memory bandwidth is not measured model speed; verify runtime and checkpoint compatibility. Hardware specification reviewed: 2026-10-13.
How estimates work
- Candidates
- Reviewed pages use published parameter counts; automatic text candidates use official repository tensor counts (packed quantized weights and unsupported architectures excluded).
- Planning
- Q4 weights + 20% allowance + reserve: 25% of unified or system memory (at least 2 GB, at most 8 GB; 6 GB here) or 2 GB per dedicated GPU. Adjustable in “Adjust the estimate”. No KV cache is calculated from context.
- Fit tiers
- Runs well: the estimate fits the fast memory (VRAM, or unified memory within the macOS GPU share). Runs slowly: within the machine's memory but outside the fast part; a graphics card offloads layers to PC RAM (the listed RAM, or 32 GB assumed). Doesn't run: above both. Tiers are planning estimates, never a speed.
- Vision
- Vision-language models use the Q4 planning estimate from the parameter count (published by the model page, or read from Hugging Face). Classic vision models (classify, detect, segment) use a 16-bit estimate. No vision benchmark is tracked, so rows carry no score.
- Media and embeddings
- Media rows use documented execution paths (exact hardware or platform) and publisher minimums; a published requirement above this memory moves a model to the next tier. Without either, a 16-bit estimate from the published parameter count (× 1.2 + 2 GB), labelled as such. Embeddings use 8-bit weights + 20% + 2 GB.
Local AI basics: parameters, Q4 and memory · Full model catalog · Data and methodology
Alternatives
Other machines on the same memory ladder (unified memory); each page lists the models that fit it.
Also 24 GB 8
- Mac mini M2 · 24 GB unified
- Mac mini M4 · 24 GB unified
- Mac mini M4 Pro · 24 GB unified
- Mac mini M5 Pro · 15-core CPU / 16-core GPU · 24 GB unified
- Mac mini M6 · 24 GB unified
- MacBook Air 13-inch M4 · 24 GB unified
- MacBook Air 13-inch M5 · 24 GB unified
- MacBook Pro 14-inch · M5 Pro · 18-core CPU / 20-core GPU · 24 GB unified
Less memory · 16 GB 8
- iMac 24-inch · M4 · 8-core CPU / 8-core GPU · 16 GB unified
- Mac mini M1 · 16 GB unified
- Mac mini M2 · 16 GB unified
- Mac mini M4 · 16 GB unified
- Mac mini M6 · 16 GB unified
- MacBook Air 13-inch M4 · 16 GB unified
- MacBook Air 13-inch M5 · 16 GB unified
- MacBook Air 15-inch · M5 · 10-core CPU / 10-core GPU · 16 GB unified
More memory · 32 GB 8
- Dell XPS 14 (2026, DA14260) · Core Ultra 7 358H · Intel Arc Graphics · 32 GB
- Mac mini M4 · 32 GB unified
- Mac mini M6 · 32 GB unified
- MacBook Air 13-inch M4 · 32 GB unified
- MacBook Air 13-inch M5 · 32 GB unified
- MacBook Air 15-inch · M5 · 10-core CPU / 10-core GPU · 32 GB unified
- MacBook Pro 14-inch (M5) · M5 · 10-core CPU / 10-core GPU · 32 GB unified
- ThinkPad T14 Gen 5 AMD · MTM 21MDS2H10D
