SWE-INTERACT: Reimagining SWE Benchmarks as User-Driven Long-Horizon Coding Sessions

2026-06-29

AI Summary

Existing software engineering benchmarks predominantly rely on static, fully specified requirements, making them inadequate for evaluating agents’ ability to handle ambiguous and dynamically evolving requirements in real-world scenarios. This work proposes the first evaluation framework supporting multi-turn interactions, leveraging a user simulator grounded in real-world coding interaction data to progressively reveal requirements, inject dynamic constraints, and provide feedback. The framework assesses agents’ capabilities in intent understanding, iterative adaptation, and sustained development. Experimental results demonstrate that while state-of-the-art models achieve a task completion rate of approximately 50% in single-turn settings, their performance drops sharply to 25% in interactive scenarios, revealing significant deficiencies in requirement tracking and robust iterative refinement.

Abstract

We introduce SWE-Interact, a new testbed for evaluating coding agents on multi-turn, interactive, user-driven software engineering tasks. Existing frontier SWE benchmarks typically provide complete requirements upfront and evaluate agents on autonomous implementation. In contrast, SWE-Interact places agents in a realistic developer workflow: a carefully designed user simulator starts with vague or incomplete instructions, progressively reveals requirements, inspects the agent's workspace, and provides targeted feedback, revisions, and new constraints until the full task goal has been handed off. Grounded in large-scale studies of real coding-agent interactions, this setup tests whether agents can discover user intent, adapt to evolving requirements, and build on their own prior work. Across a suite of frontier and open-weight models, we find that strong performance on single-turn SWE tasks does not reliably transfer to multi-turn, user-driven workflows: the best-performing models solve roughly 50% of single-turn baseline tasks but only 25% of the corresponding SWE-Interact tasks. The strongest models in our evaluation, including Opus 4.8 and GPT 5.5, start strong even in the face of vague initial instructions, persevere until all the requirements are surfaced by the user, integrate them better and write clean code. However, they still suffer from over-agentic coding, forgetting requirements and technical mistakes. Weaker models start poorly under ambiguity, give up early, forget or ignore instructions and rework their code more. Overall, SWE-Interact measures an orthogonal, real-world capability axis for frontier model development: interactive goal discovery and iterative refinement with a user in the loop.

Problem

• software engineering
• interactive coding
• user-driven tasks
• requirement evolution
• agent evaluation

Innovation

• interactive coding
• user-driven development
• long-horizon tasks
• requirement evolution
• agent evaluation benchmark