What AI models can the Apple iPhone 16 Pro / 16 Pro Max run?

RAM: 8.0 GB
Storage: 128 GB – 1 TB
One app can use: 4.4 GB
Memory bandwidth: 60 GB/s
FP16 compute: 2.2 TFLOPS
Released: 2024

176 / 622 LLMs run Best: Nemotron Labs Dory 1.0 4B 7 / 61 Image models run Best: SDXL Turbo 0 / 43 Video models run None fit 0 / 70 Decision models run No iOS app

Every iPhone 16 Pro / 16 Pro Max has 8 GB of RAM, whatever its storage (128 GB, 256 GB, 512 GB or 1 TB). A model runs in RAM (the 8.0 GB here), not in storage. Storage (128 GB, 256 GB, 512 GB or 1 TB on this version) only keeps the downloaded files, and even the smallest size has room for several. iOS lets one app hold only part of the RAM before it closes the app, so we plan around 4.4 GB; a model bigger than that won’t run. Phone apps such as PocketPal AI download a model once, then run it with no connection.

More about this hardware

Maker	Apple
Type	Mobile device (system on a chip)
App share of memory	55% (what one app can typically hold before the system closes it)
Chip	Apple A18 Pro
RAM type	8 GB LPDDR5X-7500 (teardown)
Versions sold	16 Pro: 8 GB RAM with 128 GB, 256 GB, 512 GB or 1 TB 16 Pro Max: 8 GB RAM with 256 GB, 512 GB or 1 TB
Bandwidth figure	60.0 GB/s, from teardowns (the maker publishes no memory speed)
Note	A18 Pro with a 6-core GPU; 8 GB LPDDR5X-7500 (60 GB/s).
Maker’s specifications	www.apple.com

Apps that run models well on the Apple iPhone 16 Pro / 16 Pro Max

iOS

PocketPal AIRuns the model on the device itself, offline: download it once in the app.

For a typical model that fits in memory. Each model page picks the best app for that model on this hardware. All apps

Best LLMs for the Apple iPhone 16 Pro / 16 Pro Max

Quality	Model	Quant	Memory	Speed	Runs
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.	Nemotron Labs Dory 1.0 4BNVIDIA	`Q4_K_M`	3.5 GB	19.0 tok/s	Runs well
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.	Gemma 4 E2BGoogle	`Q4_K_M`	3.8 GB	32.9 tok/s	Runs welltight fit
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.	Spark-X2.6 4BSparkLLM	`Q4_K_M`	3.4 GB	18.4 tok/s	Runs well
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	MiniCPM5 2BOpenBMB	`Q4_K_M`	2.5 GB	30.0 tok/s	Runs great
57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	Nanbeige4.2 3BBOSS Zhipin (Nanbeige)	`Q3_K_M`	4.1 GB	23.5 tok/s	Runs welltight fit
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	TMax 4BAi2	`Q4_K_M`	3.4 GB	18.0 tok/s	Runs well
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	Qwen3.5 4BAlibaba	`Q4_K_M`	3.7 GB	16.2 tok/s	Runs well
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.	Agents-A1 4BInternScience	`Q4_K_M`	3.6 GB	16.7 tok/s	Runs well
55Good quality (B) · 55 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.	Fara 1.5 4BMicrosoft	`Q4_K_M`	3.6 GB	16.7 tok/s	Runs well
42Fair quality (C) · 42 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.	Unlimited OCR 2 (3B-A0.6B)Baidu	`Q5_K_M`	3.5 GB	76.4 tok/s	Runs great
49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.	Nemotron 3.5 Bolt E4BNVIDIA	`Q2_K`	3.9 GB	27.6 tok/s	Runs welltight fit
53Fair quality (C) · 53 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.	Qwen3 4B Thinking 2507Alibaba	`Q4_K_M`	4.3 GB	18.8 tok/s	Runs welltight fit

Best Image models for the Apple iPhone 16 Pro / 16 Pro Max

Quality	Model	Quant	Memory	Speed	Runs
39Basic quality (D) · 39 of 100: how capable the model is at FP8, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.	SDXL TurboStability AI	`FP8`	4.2 GB	3.9 s	Runs welltight fit
30Basic quality (D) · 30 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.	LCM DreamShaper v7Tsinghua (LCM)	`FP16`	2.8 GB	7.2 s	Runs great
28Basic quality (D) · 28 of 100: how capable the model is at FP16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.	SD TurboStability AI	`FP16`	3.3 GB	2.3 s	Runs great
42Fair quality (C) · 42 of 100: how capable the model is at FP16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.	Realistic Vision V5.1SG161222	`FP16`	2.8 GB	35 s	Runs well
40Fair quality (C) · 40 of 100: how capable the model is at FP16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.	DreamShaper 8Lykon	`FP16`	2.8 GB	35 s	Runs well
50Fair quality (C) · 50 of 100: how capable the model is at FP16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.	PixArt-Σ XL 1024PixArt	`FP16`	3.3 GB	45 s	Slowencoder unloaded after use

Best Video models for the Apple iPhone 16 Pro / 16 Pro Max

None of the video models in our catalog fit in this much memory.

Best Decision models for the Apple iPhone 16 Pro / 16 Pro Max

None of the decision models in our catalog fit in this much memory.

Best LLMs by task on this hardware Open in Model Finder

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.