SWE-Bench Mobile (Codex): leaderboard
Metric: Task success rate (%) on SWE-Bench Mobile's 50 industrial iOS feature tasks (PRD, Figma design and a mixed Swift/Objective-C production codebase), the agent run under Codex: a task counts only when every test of its pytest suite passes on the submitted diff, a missing patch counts as failed; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 10 | #91 |
| 2 | Claude Sonnet 4.5 | 10 | #138 |
| 3 | Claude Opus 4.5 | 4 | #79 |
| 4 | GPT-5.1 | 0 | #131 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=swe-bench-mobile-codex · How It Works · Data refreshed daily, snapshot 2026-10-11.