Devin SWE-2 Max Builds an iOS App Running Gemma 4 E2B On-Device

Melvin Vivas · X video post · 2026-09-19 · 0:55 · 2,450 views · Open on X

Topics: AI Dev Tools & Productivity, AI Agents, Tool Use & MCP, Portfolio Projects · Level: beginner

Summary

Melvin Vivas quote-posts a demo where Devin, Cognition's AI coding agent, used its SWE-2 Max model to build an iOS app that runs Google's small Gemma 4 E2B model locally on an iPhone. He suggests Codex users try it. The post says SWE-2 Max is free for now. The video has no narration, so the lesson is what the demo shows: a coding agent can build a working on-device LLM mobile app.

Key points

• Devin (Cognition's autonomous coding agent) was asked to build a full iOS app from a prompt.
• The app runs the Gemma 4 E2B model locally on the iPhone, with no cloud inference.
• The build used the SWE-2 Max model in Devin, which the post says was free at the time (Sept 2026).
• Gemma 'E2B' variants are small, edge-sized models built to run on phones and other low-resource devices.
• The creator frames Devin SWE-2 as an alternative worth trying for people who use OpenAI Codex.
• The clip has no speech, so it shows no step-by-step build process.

Resources mentioned

• Devin · tool · x.com · paid
  A desktop app from the makers of the Devin coding agent for planning, delegating, reviewing and shipping work across fleets of local and cloud coding agents.
  Also in: Using the Devin iOS App to Build iOS Apps (Quick Demo) (Melvin Vivas on X · notes), Use Your ChatGPT Plus/Pro Subscription to Run Devin (Melvin Vivas on X · notes), OpenAI DevDay recap: Dots, GPT-6.1 Sol, Codex Cloud, Agents API (Melvin Vivas on X · notes), Devin price cuts and top score on FrontierCode 1.1 Extended (Melvin Vivas on X · notes) and 44 more
• SWE-2 · tool · cognition.com · free
  Cognition's SWE-2 coding model (Max tier) that runs inside Devin for agentic software tasks.
  Also in: Comparing Coding Models on the Same Task in Devin iOS (Melvin Vivas on X · notes), Devin's SWE-2 Coding Model Is Free for a Limited Time (Until Oct 8/15) (Melvin Vivas on X · notes), Devin's SWE-2 model is free on paid plans for a limited time (Melvin Vivas on X · notes), Using Devin CLI with the free SWE-2 model for sysadmin tasks (Melvin Vivas on X · notes) and 8 more
• Gemma 4 · tool · ai.google.dev · free
  Google's family of open-weight models in several sizes, built to run on devices and offline, with multimodal and agentic abilities, and open to fine-tuning.
  Also in: Gemma 4 Runs Locally On-Device in the Antigravity SDK (Melvin Vivas on X · notes), On-Device AI: Running Gemma 4 E2B Offline on an iPhone with LiteRT (Melvin Vivas on X · notes), Running Gemma 4 Models Offline on an iPhone (Melvin Vivas on X · notes), Fine-tuning Gemma4-E2B on your own tweet style with Unsloth (Melvin Vivas on X · notes) and 25 more
• OpenAI Codex · tool · openai.com · paid
  OpenAI's coding agent. In the diagram it writes code, fixes review findings and drives the build loop. The creator also used it to make this video.
  Also in: An agent bot that installs and drives Codex on its own (Melvin Vivas on X · notes), Asking a Coder bot to install Codex (Melvin Vivas on X · notes), Sign in with ChatGPT: Setting Usage Limits for Each App (Melvin Vivas on X · notes), Codex Cloud Environments Must Be Saved & Published Before Use (Melvin Vivas on X · notes) and 240 more

Try this

• Try Devin's SWE-2 Max while it is free, especially if you use Codex now.
• Compare how Devin and Codex handle the same end-to-end app-building task.
• Use an AI coding agent to build an iOS app that runs a small open model like Gemma 4 E2B fully on the iPhone, with no cloud inference.