Stack Overflow 2025 Developer Survey

The 2025 Developer Survey from Stack Overflow gathered input from over 49,000 developers in 177 countries. The survey covered 62 questions on 314 technologies, with new focus areas including AI agent tools, large language models (LLMs), and developer community platforms. Now in its fifteenth year, it provides a broad look at the state of software development worldwide.

Stack Overflow 2025 Developer Survey results: Most admired and desired technologies

In cloud development and infrastructure tools, Cargo ranked as the most admired, with 71% of respondents rating it highly. Terraform, Homebrew, Make, and APT also scored well in admiration and desire among developers. Rust’s continued rise is tied closely to the popularity of Cargo.

Among large language models, Anthropic’s Claude Sonnet was the most admired, with OpenAI GPT models as the most desired. Gemini Reasoning and Gemini Flash also appeared prominently in the results.

AI tools remain a significant part of development workflows, with 84% of respondents either using or planning to use them. Daily usage was reported by 47% of developers, and 51% of professional developers use AI tools every day. Despite this, trust remains limited, with only 3% highly trusting AI outputs and 46% expressing some level of distrust.

AI agents have yet to see mainstream adoption, with 38% of developers stating they have no plans to use them. However, among those who do, 69% report increased productivity and 70% say agents reduce time on tasks. The main frustrations with AI tools include solutions that are almost correct but still require time-consuming debugging.

Work settings vary, with 32% working fully remote and 18% in-person. Hybrid models are common, with variations leaning either toward flexibility or in-office presence. The United States leads in remote work, with 45% of developers working from home full-time. Job satisfaction has improved, with 24% happy in their roles compared to 20% last year.

Python adoption grew by 7 percentage points over last year, now used by 58% of developers. JavaScript, HTML/CSS, and SQL remain widely used, while Bash/Shell also holds a strong presence. OpenAI GPT models are the most popular LLMs, used by 81% of respondents. Claude Sonnet saw stronger adoption among professional developers compared to those still learning.

Visual Studio Code and Visual Studio remain the top development environments for the fourth year. GitHub replaced Jira as the most desired collaboration tool, and Markdown files maintained their position as the most admired documentation tool. Stack Overflow tags such as “uv” also drew attention, with uv becoming the most admired in its category.

Stack Overflow remains a primary resource, with 82% of developers visiting at least a few times per month and 25% visiting daily or more. While lists and articles remain popular content formats, younger developers show stronger interest in interactive formats like coding challenges and live chat.

YouTube stands out as a key community platform for those learning to code, surpassing its use among professional developers. GitHub, Reddit, and Stack Exchange also maintain strong engagement.

The most common reasons developers stop using a technology are security or privacy concerns, high costs, and better alternatives. Usability, inefficiency, and outdated features follow. The lack of AI integration ranks low as a deal-breaker.

Respondents come from diverse backgrounds, with the USA, Germany, and India as the top three countries. Over a third of developers have been coding for less than 10 years, and a notable portion have 21 or more years of experience. Full-stack developer remains the top role, with architect roles debuting at fourth place.

More than one third of respondents learned AI-enabled tools for work in the past year, and 31% learned them for personal interest. While AI tools are widely used, most developers recognize the need for human verification and prefer to understand the code they work with.

When using AI agents, developers report benefits in productivity and learning, but limited improvements in team collaboration. The data suggests AI is more valuable for personal efficiency than for group work. Subscription-based AI IDEs have not overtaken established tools, showing that extensions and add-ons remain the preferred way to integrate AI into development environments.

2025 Developer Survey results