On this page
You can run Qwen3-VL on your iPhone with Off Grid AI Mobile (OGAM). Search for Qwen3-VL, download Qwen’s own 2B or 4B build, and your iPhone reads receipts, letters and screenshots with no internet and no uploads.
It’s free, there’s no account, and the 2B fits iPhones with as little as 4 GB of RAM.
App Store · Download Off Grid AI
What Qwen3-VL is
Qwen3-VL is Alibaba’s vision-language family. Qwen released it in October 2025 under the Apache 2.0 licence and published its own GGUF files, the format OGAM runs. The phone sizes are 2B, 4B and 8B.
Qwen’s model card lists what’s new over the older Qwen2.5-VL. The parts that matter on a phone:
- Reads text in 32 languages, up from 19, and copes with low light, blur and tilted pages.
- Understands app screens: it picks out buttons and what they do.
- Finds things in a picture and says where they sit.
Which size fits your iPhone
OGAM decides fit from your iPhone’s RAM. A download under 60% of RAM runs well, 60% to 70% is marked Tight, and anything bigger isn’t offered. The download is the model file plus its vision file, because the app fetches both.
| Model | Model file (Q4_K_M) | Vision file | Total | 6 GB iPhone | 8 GB iPhone | 12 GB iPhone |
|---|---|---|---|---|---|---|
| Qwen3-VL 2B | 1.11 GB | 0.82 GB | 1.9 GB | Runs well | Runs well | Runs well |
| Qwen3-VL 4B | 2.50 GB | 0.84 GB | 3.3 GB | Runs well, just | Runs well | Runs well |
| Qwen3-VL 8B | 5.03 GB | 1.16 GB | 6.2 GB | No | No | Runs well |
File sizes come from Qwen’s GGUF repos. iOS reports a little less RAM than the box says, so a 12 GB iPhone 17 Pro Max shows 11.2 GB, and the limits sit slightly below the table. That’s why the 4B can show Tight on a 6 GB iPhone.
As a rule of thumb, the iPhone 15 Pro and later have 8 GB or more, so pick the 4B there. On an older iPhone, pick the 2B.
Set it up
- Install OGAM from the App Store.
- Tap Models and stay on the Text tab.
- Type Qwen3-VL in Search models.
- Tap Qwen3-VL-4B-Instruct-GGUF, or the 2B or 8B. Qwen’s own upload leads each card, and other uploads of the same model are counted in a +N uploads chip.
- Under Sizes, take the Recommended file and download it. Use Wi-Fi, because the 4B is about 3.3 GB with its vision file.
- When it finishes, tap Chat now.
The cards say Newer: Qwen 3.5. That’s OGAM pointing you at the model it recommends today. It doesn’t stop you, and Qwen3-VL downloads and runs like any other model.
Once it’s downloaded, turn on airplane mode and ask again. It still answers, because the model runs on your iPhone.
Ask it about a photo
Tap + in the message box, choose Photo, then Camera or Photo Library. Add your question and send.
Reading a receipt photo (iPhone)
|
Things to ask:
- A receipt: “List each item with its price, then the total.”
- A letter in Spanish or Hindi: “What is this letter asking me to do, and by when?”
- A screenshot of a settings page: “Which switch turns off location sharing?”
- A photo of a desk: “Where are my keys?”
Crop to the part that matters before you send. A tight photo of one receipt reads better than a whole table.
Qwen3-VL or Qwen 3.5?
Qwen 3.5 is newer and also reads images, so OGAM lists it first. Pick Qwen3-VL when your photos are mostly text in languages other than English, or when you want a 4B vision model to sit between Qwen 3.5’s 2B and 9B. If you want the app’s own pick instead, the Qwen 3.5 iPhone guide covers it.
If something goes wrong
| Symptom | Check | Fix |
|---|---|---|
| Qwen3-VL 8B is missing, or has no size to download | Your iPhone has less than 12 GB of RAM | Pick the 4B |
| A size is marked Tight | Other apps open, long chats | Close other apps, or pick the next size down |
| Photo says “Vision Not Supported” | The active model reads text only | Switch to the Qwen3-VL model you downloaded |
| The app says “Vision File Missing” | The model is there, its vision file isn’t | Open your downloads from that message, tap the wrench next to the model, then reload it |
| Answers miss small print | The photo is too wide or blurry | Crop closer and retake in better light |
FAQ
Can Qwen3-VL run on an iPhone? Yes. Search for Qwen3-VL in OGAM and download Qwen’s own GGUF upload. The app adds the vision file, and the model reads photos on the iPhone with no internet. The 2B fits iPhones with 4 GB of RAM, the 4B wants 6 to 8 GB and the 8B needs a 12 GB iPhone.
Which Qwen3-VL size should I pick for my iPhone? The 4B on an 8 GB iPhone (iPhone 15 Pro or later). The 2B on older iPhones. The 8B only on a 12 GB iPhone such as the iPhone 17 Pro Max, because its 4-bit file plus vision file is about 6.2 GB.
Why does the app say Newer: Qwen 3.5 on Qwen3-VL? OGAM now recommends Qwen 3.5, which also reads images. The label is a heads-up, not a block. Qwen3-VL still downloads and runs, so you can pick it if you want it.
Can Qwen3-VL read documents in other languages? Qwen says Qwen3-VL reads text in 32 languages, and it holds up in low light, blur and tilt. That makes it a good pick for photos of letters, menus and forms that aren’t in English.
Is it free and private? Yes. OGAM is free with no account. After the download, the model runs in your iPhone’s memory, so your photos aren’t uploaded to any server.
Related guides
- How to Run Vision AI Locally on Your iPhone in 2026 (Completely Offline, No Account)
- How to Ask AI About a Screenshot on iPhone in 2026 Without Internet
- How to Find AI Models Your iPhone Can Actually Run in 2026 (No Failed Downloads)
- How to Run Qwen 3.5 on Your iPhone in 2026 (Completely Offline, No Subscription)
Get Off Grid AI on the App Store and point Qwen3-VL at your next receipt.