Apple M3 Pro
36 GB decides what fits. 150 GB/s decides how fast it runs once it does.
Computed for 36 GB, the largest configuration. The Apple M3 Pro is also sold with 18 GB, and what fits changes with it.
Find one: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
- Largest popular model that fitsKimi-Linear-48B-A3B-Instruct49B params · Q4_K_M · needs 30.4 GB~53 tok/s Faster than you readOpen in the calculator →
- Most downloaded that fitsQwen3.8-27B28B params · Q4_K_M · needs 18.6 GB~9.0 tok/s SlowOpen in the calculator →
At 8,192 tokens of context with the whole model in device memory. Speeds are estimates from memory bandwidth, not benchmarks run on this card.
217 of 320 fit entirely
217 of the 320 models the engine can size fit entirely in device memory at Q4_K_M where it is published, otherwise the nearest published format, and 8,192 tokens. Unified memory is one pool, so there is no second memory tier to spill into; the remaining 103 do not run at this context.
Featured models · Q4_K_M at 8,192 tokens, including offload
| Model | Needs | Verdict | Decode | Calculator |
|---|---|---|---|---|
| Qwen3.5-2B2.3B parameters | 2.32 GB | fits | ~69 tok/s | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF164.0B parameters | 3.46 GB | fits | ~41 tok/s | Open → |
| Qwen3-VL-8B-Instruct8.8B parameters | 7.04 GB | fits | ~21 tok/sAbout reading pace | Open → |
| Qwen3.5-9B9.7B parameters | 7.05 GB | fits | ~25 tok/sAbout reading pace | Open → |
| gemma-4-12B-it12B parameters | 9.54 GB | fits | ~16 tok/sAbout reading pace | Open → |
| gpt-oss-20b21B parameters | 13.8 GB | fits | ~43 tok/s | Open → |
| Qwen3.8-27B28B parameters | 18.6 GB | fits | ~9.0 tok/sSlow | Open → |
| Qwen3.6-27B28B parameters | 18.6 GB | fits | ~9.0 tok/sSlow | Open → |
| Kimi-Linear-48B-A3B-Instruct49B parameters | 30.4 GB | fits | ~53 tok/s | Open → |
| Qwen3.8-Flash-Next180B parameters | 111.6 GB | short by 75.6 GB | does not run | Open → |
| DeepSeek-V4-Flash-Vision-Exp305B parameters | 188.1 GB | short by 152.1 GB | does not run | Open → |
| GLM-5.3-Flash321B parameters | 198.3 GB | short by 162.3 GB | does not run | Open → |
These examples are selected from prominent labs using the catalogue’s latest Hugging Face 30-day downloads and repository-creation freshness signal, with newer releases guaranteed a place. Offloaded rows assume 32 GB of system RAM, and a speed is only shown for a row that runs. Fitting in memory is not the same as loading: whether the runtime and version you have supports each architecture and format on this machine has not been tested here. Each “Open” link carries the same model, format, context and RAM into the calculator.
Fits entirely in Apple M3 Pro (36 GB) memory — 217 of 320 sized models
The most demanding model that fits is Kimi-Linear-48B-A3B-Instruct at Q4_K_M: 30.4 GB of the 36 GB, leaving 5.63 GB spare.
Fully resident at 8,192 tokens (or the model’s own maximum, where that is shorter), offload off, at 150 GB/s. Each row is a run of the engine for this configuration; the rows start with current, prominent releases and “fits” is memory, not a tested runtime. Older or less prominent models remain available through this search and “Show all”.
Showing 40 of 217 models that fit.
| Model | Format | Needs | Spare | Decode | Downloads / 30d | Calculator |
|---|---|---|---|---|---|---|
| Qwen3.8-27BQwen · 28B params | Q4_K_M | 18.6 GB | 17.4 GB | ~9.0 tok/sSlow | 6.9M | Open → |
| gemma-4-26B-A4B-itGoogle · 26B params | Q4_K_M | 16.8 GB | 19.2 GB | ~41 tok/s | 13M | Open → |
| gemma-4-31B-itGoogle · 31B params | Q4_K_M | 22.2 GB | 13.8 GB | ~6.6 tok/sSlow | 9.9M | Open → |
| Qwen3.5-9BQwen · 9.7B params | Q4_K_M | 7.05 GB | 29.0 GB | ~25 tok/sAbout reading pace | 9M | Open → |
| Qwen3.5-4BQwen · 4.7B params | Q4_K_M | 3.98 GB | 32.0 GB | ~40 tok/s | 7.8M | Open → |
| Qwen3.6-35B-A3BQwen · 36B params | Q4_K_M | 23.1 GB | 12.9 GB | ~53 tok/s | 3.3M | Open → |
| gemma-4-12B-itGoogle · 12B params | Q4_K_M | 9.54 GB | 26.5 GB | ~16 tok/sAbout reading pace | 1.9M | Open → |
| Qwen3.6-27BQwen · 28B params | Q4_K_M | 18.6 GB | 17.4 GB | ~9.0 tok/sSlow | 2.5M | Open → |
| Qwen3.5-2BQwen · 2.3B params | Q4_K_M | 2.32 GB | 33.7 GB | ~69 tok/s | 4.9M | Open → |
| gemma-4-E4B-itGoogle · 8.0B params | Q4_K_M | 5.86 GB | 30.1 GB | ~37 tok/s | 4.4M | Open → |
| NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA · 32B params | Q4_K_M | 20.3 GB | 15.7 GB | ~6.9 tok/sSlow | 530K | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA · 4.0B params | Q4_K_M | 3.46 GB | 32.5 GB | ~41 tok/s | 3.4M | Open → |
| gemma-4-E2B-itGoogle · 5.1B params | Q4_K_M | 4.02 GB | 32.0 GB | ~72 tok/s | 3M | Open → |
| Qwen3.5-0.8BQwen · 873M params | Q4_K_M | 1.46 GB | 34.5 GB | ~103 tok/s | 2.6M | Open → |
| Qwen3-VL-8B-InstructQwen · 8.8B params | Q4_K_M | 7.04 GB | 29.0 GB | ~21 tok/sAbout reading pace | 14.6M | Open → |
| Qwen3.5-27BQwen · 28B params | Q4_K_M | 18.6 GB | 17.4 GB | ~9.0 tok/sSlow | 1.9M | Open → |
| North-Micro-Vision-InstructCohereLabs · 2.5B params | Q4_K_M | 2.99 GB | 33.0 GB | ~53 tok/s | 180.7K | Open → |
| Qwen3.5-35B-A3BQwen · 36B params | Q4_K_M | 23.1 GB | 12.9 GB | ~53 tok/s | 1.6M | Open → |
| granite-4.2-8bIBM · 8.8B params | Q4_K_M | 7.49 GB | 28.5 GB | ~19 tok/sAbout reading pace | 128.4K | Open → |
| GLM-4.7-Flashzai-org · 31B params | Q4_K_M | 20.4 GB | 15.6 GB | ~6.9 tok/sSlow | 1.8M | Open → |
| granite-4.1-3bIBM · 3.4B params | Q4_K_M | 3.72 GB | 32.3 GB | ~39 tok/s | 521.8K | Open → |
| LFM2.5-2.6BLiquidAI · 2.7B params | Q4_K_M | 2.61 GB | 33.4 GB | ~56 tok/s | 108.1K | Open → |
| Qwen3-VL-4B-InstructQwen · 4.4B params | Q4_K_M | 4.72 GB | 31.3 GB | ~31 tok/s | 3.5M | Open → |
| granite-4.1-30bIBM · 29B params | Q4_K_M | 20.4 GB | 15.6 GB | ~6.9 tok/sSlow | 301.5K | Open → |
| Qwen3-VL-2B-InstructQwen · 2.1B params | Q4_K_M | 3.05 GB | 33.0 GB | ~49 tok/s | 2.8M | Open → |
| granite-4.2-3bIBM · 3.7B params | Q4_K_M | 3.72 GB | 32.3 GB | ~39 tok/s | 50.6K | Open → |
| LFM2.5-230MLiquidAI · 230M params | Q4_K_M | 1.05 GB | 34.9 GB | ~134 tok/s | 88.6K | Open → |
| Qwen3-0.6BQwen · 752M params | Q4_K_M | 2.22 GB | 33.8 GB | ~66 tok/s | 29.7M | Open → |
| gpt-oss-20bOpenAI · 21B params | Q4_K_M | 13.8 GB | 22.2 GB | ~43 tok/s | 6.6M | Open → |
| granite-4.2-30bIBM · 29B params | Q4_K_M | 20.7 GB | 15.3 GB | ~6.9 tok/sSlow | 35.8K | Open → |
| granite-4.1-8bIBM · 8.8B params | Q4_K_M | 7.49 GB | 28.5 GB | ~19 tok/sAbout reading pace | 179.3K | Open → |
| NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA · 32B params | Q4_K_M | 20.3 GB | 15.7 GB | ~6.9 tok/sSlow | 875.2K | Open → |
| LFM2.5-VL-3BLiquidAI · 3.1B params | Q4_K_M | 2.85 GB | 33.1 GB | ~56 tok/s | 26.7K | Open → |
| MiMo-V2.6-Distill-Qwen-9BNEWXiaomiMiMo · 9.4B params | Q4_K_M | 6.90 GB | 29.1 GB | ~25 tok/sAbout reading pace | 14.6K | Open → |
| Qwen3-4B-Instruct-2507Qwen · 4.0B params | Q4_K_M | 4.72 GB | 31.3 GB | ~31 tok/s | 3.7M | Open → |
| Ling-3.0-tinyinclusionAI · 7.9B params | Q4_K_M | 5.85 GB | 30.1 GB | ~85 tok/s | 17.7K | Open → |
| Qwen3-8BQwen · 8.2B params | Q4_K_M | 7.04 GB | 29.0 GB | ~21 tok/sAbout reading pace | 10.7M | Open → |
| Olmo-3-7B-Instructallenai · 7.3B params | Q4_K_M | 7.96 GB | 28.0 GB | ~18 tok/sAbout reading pace | 481.7K | Open → |
| Qwen3-4BQwen · 4.0B params | Q4_K_M | 4.51 GB | 31.5 GB | ~31 tok/s | 7.8M | Open → |
| LFM2.5-8B-A1BLiquidAI · 8.5B params | Q4_K_M | 6.06 GB | 29.9 GB | ~74 tok/s | 32.4K | Open → |
Sized at the format most people actually download, not at FP16. The catalogue ordered by downloads →
Find the Apple M3 Pro: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
The specification behind every figure
What the manufacturer publishes for this device, and the pages it was read from.
Manufacturer specification
| Memory scope | unified-system |
|---|---|
| Capacity | 36 GB |
| Published options | 18 GB, 36 GB |
| Bandwidth | 150 GB/s |
| Memory type | unified memory |
| Bus width | Not published |
| FP32 peak | Not published |
| Dense matrix peak | Not published without sparsity |
| Power | Not published |
Source ledger
Caveats
- Unified memory is shared with the operating system and applications; their use is not subtracted.
- Apple publishes no peak throughput figure for this chip, so no compute roof is priced for it and time to first token is withheld.
Published capacity is a hardware ceiling, not guaranteed free runtime memory. The calculator shows the runtime reserve separately rather than folding it into a single number.
Run the diagnostic on the Apple M3 Pro →