Software Engineering · arXiv:2602.09540 v1 · submitted 10 Feb 2026
SWE-Bench Mobile: Can Large Language Model Agents Develop Industry-Level Mobile Applications?
The paper says
“These findings highlight a significant gap between current agent capabilities and industrial requirements, while providing actionable insights for practitioners and researchers.”
from the abstract · picked out as the main claim by a classifier
Reviews
Reviews by people
No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
Can large language model agents develop industry-level mobile applications? We introduce \textbfSWE-Bench Mobile, a benchmark for evaluating coding agents on realistic software engineering tasks derived from a production iOS codebase. Unlike existing benchmarks that focus on isolated problems or bug fixes, SWE-Bench Mobile captures the full complexity of industrial development: multi-modal inputs (PRDs and Figma designs), a large-scale mixed Swift/Objective-C codebase, and comprehensive test suites. We evaluate 22 agent-model configurations across four coding agents – three commercial (Cursor, Codex, Claude Code) and one open-source (OpenCode) – and find that even the best configurations achieve only 12% task success rate. Our analysis reveals that (1) agent design matters as much as model capability – the same model shows up to 6× performance gap across agents, (2) commercial agents consistently outperform open-source alternatives, and (3) simple “Defensive Programming” prompts outperform complex ones by 7.4%. These findings highlight a significant gap between current agent capabilities and industrial requirements, while providing actionable insights for practitioners and researchers. We release SWE-Bench Mobile as a hosted benchmark challenge to prevent data contamination and ensure fair evaluation. The public leaderboard and development toolkit are available at https://swebenchmobile.com.
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-09-23 · no language-model review
How the main claim is stated: asserted 0.98
What the abstract itself states, as probabilities: claims first 0.13 · states limitation 0.67 · human participants 0.06 · controlled comparison 0.30 · randomized assignment 0.02 · states quantitative result 0.99 · compares with prior results 0.76 · mentions code or data availability 0.95
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| Can large language model agents develop industry-level mobile applications? | background 0.91 | 0.04 | 0.02 | unclear 0.96 |
| We introduce \textbf{SWE-Bench Mobile}, a benchmark for evaluating coding agents on realistic software engineering tasks derived from a production iOS codebase. | contribution 1.00 | 0.93 | 0.04 | asserted 1.00 |
| Unlike existing benchmarks that focus on isolated problems or bug fixes, SWE-Bench Mobile captures the full complexity of industrial development: multi-modal inputs (PRDs and Figma designs), a large-scale mixed Swift/Objective-C codebase, and comprehensive test suites. | contribution 0.80 | 0.54 | 0.03 | asserted 1.00 |
| We evaluate 22 agent-model configurations across four coding agents -- three commercial (Cursor, Codex, Claude Code) and one open-source (OpenCode) -- and find that even the best configurations achieve only 12\% task success rate. | result 1.00 | 0.98 | 0.48 | asserted 1.00 |
| Our analysis reveals that (1) agent design matters as much as model capability -- the same model shows up to 6$\times$ performance gap across agents, (2) commercial agents consistently outperform open-source alternatives, and (3) simple ``Defensive Programming'' prompts outperform complex ones by 7.4\%. | result 0.95 | 0.97 | 0.37 | asserted 1.00 |
| These findings highlight a significant gap between current agent capabilities and industrial requirements, while providing actionable insights for practitioners and researchers. | interpretation 1.00 | 0.60 | 0.68 | asserted 0.98 |
| We release SWE-Bench Mobile as a \textit{hosted benchmark challenge} to prevent data contamination and ensure fair evaluation. | contribution 0.80 | 0.90 | 0.03 | asserted 0.99 |
| The public leaderboard and development toolkit are available at https://swebenchmobile.com. | contribution 0.95 | 0.62 | 0.02 | asserted 1.00 |
What kind of paper the classifier reports this is: dataset or benchmark 1.00 · evidence from observation or experiment 1.00
What was examined, and what was not
Sources read
- arXiv listing abstract v1
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 6,374 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
Cited by 2 Pith papers
Reviewed papers on Pith that cite this one. The full citation record
- LiveUIBench: Evaluating Agent as App Workflows cs.SE · Oct 2026
- HASTE: Evolving Agent Harnesses Against Emerging Attacks Using Sparse Evidence cs.AI · Oct 2026
How the literature uses it
7 arXiv papers cite this in 10 sentences
Papers citing it · 6 sentences without a confident label
Uses its method or tool 1 paper
-
Strategy B, Diff-Based Intent, inspired by SWE-Bench Mobile [ref] , verifies that the source code contains the expected fix pattern rather than testing runtime behavior.
SWE-Bench 5G: Benchmarking AI Coding Agents on Telecom Network Engineering Tasks · Dual Test Strategy · 0.97
Cites it as support 1 paper
-
These systems shift memory from external knowledge access to persistent experience accumulation [ref] .
From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI · Tool Invocation and Core Agent Capabilities · 0.97
Cites it as background 2 papers
-
SWE-Bench Mobile [ref] moved to iOS development with a 500K-line Swift/Objective-C codebase, introducing diff-based intent tests that verify patch correctness through structural source code inspection rather than runtime execution.
SWE-Bench 5G: Benchmarking AI Coding Agents on Telecom Network Engineering Tasks · Expanding Language and Task Scope · 0.83
-
Large language model (LLM) agents have been assisting humans across various domains, such as software engineering (SWE) and customer service [ref] .
Towards Direct Evaluation of Harness Optimizers via Priority Ranking · Introduction · 0.99
Challenges and discussion
Know something about this paper? Leave a comment.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment