AI Tool Compare

Research

AI Coding Assistant Statistics 2026

AI coding assistant statistics describe how many software developers use AI tools that suggest, write, review or debug code, which tools they use, how much they trust the output and what productivity effects have been measured. Sources include developer surveys, platform data, controlled trials and vendor announcements.

This page collects dated figures from the Stack Overflow Developer Survey (2025 and 2026 editions), Google's DORA research, the JetBrains developer ecosystem survey, GitHub's Octoverse report and one randomized trial by METR. Vendor-published figures sit in their own group and are labeled as vendor claims in each note. Pricing, benchmark scores and model rankings are not covered, and surveys of developers reflect self-selected respondents.

28 figures from 7 publishers. Each carries its source, year and a link to the page it was read on. Cite a single figure by its number, for example AI Coding Assistant Statistics 2026, figure 1.

How many developers use AI coding tools?

  1. #1
    66%

    In the 2026 Stack Overflow survey of more than 30,000 developers, 66% named coding assistants and agents as a way they use AI, ahead of general chatbots at 63%.

    Stack Overflow, 2026 Stack Overflow Developer Survey, AI section (2026). Self-selected online survey; respondents are more engaged with Stack Overflow than the average developer.

  2. #2
    73%

    Among 2026 Stack Overflow respondents who use AI coding assistants or agents, 73% use them every day.

    Stack Overflow, The results of the 2026 developer survey are here (2026). Share of users of coding assistants or agents, not of all respondents.

  3. #3
    31% for 4+ hours

    Of daily AI users in the 2026 Stack Overflow survey, 31% spend four or more hours a day working with AI, 24% two to four hours and 26% one to two.

    Stack Overflow, The results of the 2026 developer survey are here (2026). Base is respondents who use AI daily, which the survey puts above 70% of respondents.

  4. #4
    84%

    In 2025, 84% of Stack Overflow survey respondents used or planned to use AI tools in their development process, up from 76% the year before.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Includes people who only plan to use AI tools; 16.2% did not plan to.

  5. #5
    50.6%

    In 2025, 50.6% of professional developers in the Stack Overflow survey used AI tools every day, and 47.1% of all respondents did.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). A further 17.7% of respondents used AI tools weekly.

  6. #6
    90%

    Google's 2025 DORA research found that 90% of surveyed software professionals use AI, up 14% from the prior year, with a median of two hours a day.

    Google Cloud (DORA), How are developers using AI? Inside our 2025 DORA report (2025). Survey of nearly 5,000 technology professionals worldwide. Google is also a vendor of AI coding tools.

  7. #7
    85%

    JetBrains' 2025 ecosystem survey found 85% of developers regularly use AI tools for coding, and 62% rely on at least one AI assistant, agent or editor.

    JetBrains, State of Developer Ecosystem 2025 (2025). 24,534 developers in 194 countries, April to June 2025. JetBrains sells AI tools and its users may respond more.

  8. #8
    80%

    GitHub reported that roughly 80% of new developers on its platform use GitHub Copilot within their first week.

    GitHub, Octoverse: A new developer joins GitHub every second as AI leads TypeScript to #1 (2025). Vendor-published platform data from GitHub's Octoverse 2025 report; Copilot Free launched in late 2024.

Which AI coding tools and agents do developers use?

  1. #9
    66% Claude Code

    In 2026 Stack Overflow respondents, Claude Code was the most-used coding agent at 66%, followed by GitHub Copilot at 59%.

    Stack Overflow, 2026 Stack Overflow Developer Survey, AI section (2026). Share of respondents using each agent; a self-selected sample, so it is not market share.

  2. #10
    81.7% ChatGPT

    In 2025, 81.7% of Stack Overflow respondents who use AI tools had used ChatGPT and 67.9% had used GitHub Copilot, the two most common.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Google Gemini followed at 47.4%. Multiple answers were allowed.

  3. #11
    31%

    In 2025, 31% of Stack Overflow respondents used AI agents at work, with 14.1% using them daily, and 37.9% did not plan to.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Agent use was at an early stage in the 2025 wave; compare with the 2026 survey's agent figures.

  4. #12
    34% kept, 25% switched for quality

    In 2026, 34% of Stack Overflow respondents kept their original AI tool over the past year, and 25% switched for better output quality.

    Stack Overflow, 2026 Stack Overflow Developer Survey, AI section (2026). About 11% switched because an organization required a different tool, 7% for workflow integration and 6% for cost.

  5. #13
    1 million+

    GitHub's Copilot coding agent created more than one million pull requests between May and September 2025, after launching in preview in March 2025.

    GitHub, Octoverse: A new developer joins GitHub every second as AI leads TypeScript to #1 (2025). Vendor-reported count of pull requests the agent authored, not the number merged.

  6. #14
    50%

    GitHub reported in 2025 that 50% of open source projects have at least one maintainer using GitHub Copilot.

    GitHub, Octoverse: A new developer joins GitHub every second as AI leads TypeScript to #1 (2025). Vendor-published platform data; how GitHub defines a project and a maintainer is in the Octoverse report.

Do developers trust AI-generated code?

  1. #15
    33% trust, 46% distrust

    Only 33% of 2025 Stack Overflow respondents trusted the accuracy of AI output, while 46% distrusted it, and just 3.1% highly trusted it.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Trust fell as use rose; experienced developers were the most cautious.

  2. #16
    66%

    The top frustration with AI tools in 2025 was solutions that are almost right but not quite, cited by 66% of Stack Overflow respondents.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). A further 45.2% said debugging AI-generated code takes more time.

  3. #17
    45.2%

    In 2025, 45.2% of Stack Overflow respondents said debugging AI-generated code is more time-consuming, and 20% felt less confident in their own problem-solving.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Multiple answers were allowed on the frustrations question.

  4. #18
    48%

    In 2026, 48% of Stack Overflow respondents said they trust AI output when they can easily validate it, while 16% trust it for most tasks that are not important decisions.

    Stack Overflow, 2026 Stack Overflow Developer Survey, AI section (2026). The 2026 survey frames trust as conditional on verification rather than as a single score.

  5. #19
    62%

    In the 2026 survey, 62% of all respondents viewed AI at least somewhat positively, rising to 71% among daily users.

    Stack Overflow, The results of the 2026 developer survey are here (2026). The 2025 survey reported 60% favorable, down from over 70% in 2023 and 2024.

  6. #20
    24%

    Only 24% of respondents in Google's 2025 DORA survey trusted AI-generated code a great deal or a lot, while 30% trusted it a little or not at all.

    Google Cloud (DORA), How are developers using AI? Inside our 2025 DORA report (2025). Trust is reported separately from the 90% adoption figure; the rest answered a moderate amount.

  7. #21
    93%

    In the 2026 Stack Overflow survey, 93% of respondents said source attribution is required for them to trust AI output.

    Stack Overflow, The results of the 2026 developer survey are here (2026). Share of all 2026 respondents; the survey is run by Stack Overflow, a site whose content is the source of attribution.

Does AI make developers more productive?

  1. #22
    19% slower

    A randomized controlled trial of 16 experienced open-source developers found that allowing AI tools increased task completion time by 19% across 246 tasks.

    METR, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity (2025). Early-2025 tools, mostly Cursor Pro with Claude models, on mature repositories the developers knew well.

  2. #23
    24% forecast, 20% perceived

    Before the METR trial developers forecast that AI would cut completion time by 24%, and afterward they estimated it had cut time by 20%.

    METR, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity (2025). The gap between perceived and measured speed is the study's central finding; a small sample of 16 developers.

  3. #24
    80%+

    Over 80% of respondents in Google's 2025 DORA survey said AI had increased their productivity, and 59% reported a positive influence on code quality.

    Google Cloud (DORA), How are developers using AI? Inside our 2025 DORA report (2025). Self-reported perceptions, not measured output; the METR trial found a gap between the two.

  4. #25
    52%

    In 2025, 52% of Stack Overflow respondents agreed that AI tools or agents had a positive effect on their productivity, and 69% of agent users said agents raised it.

    Stack Overflow, 2025 Stack Overflow Developer Survey, AI section (2025). Self-reported; the agent figure is 68.7% of respondents who use agents.

  5. #26
    1 hour+ for nearly 9 in 10

    JetBrains found that nearly nine in ten AI users save at least one hour a week, and one in five saves eight hours or more.

    JetBrains, State of Developer Ecosystem 2025 (2025). Self-reported time savings from 24,534 developers surveyed April to June 2025.

What do vendors claim about AI-written code?

  1. #27
    $2.5 billion

    Anthropic said in February 2026 that Claude Code's run-rate revenue exceeded $2.5 billion and had more than doubled since the start of 2026.

    Anthropic, Anthropic raises $30 billion in Series G funding (2026). Vendor claim in a funding announcement, not independently audited. Weekly active users doubled since January 1, 2026.

  2. #28
    75%

    Google CEO Sundar Pichai said on April 24, 2026 that 75% of Google's new code is AI-generated, up from about 25% in 2024 and 50% in fall 2025.

    Semafor, Google CEO says 75% of company's new code is AI-generated (2026). Vendor claim by a company that sells AI coding tools; the definition of AI-generated code was not published.

Frequently asked questions

What percentage of developers use AI coding tools?

Surveys put it between 84% and 90%, depending on wording. Stack Overflow's 2025 survey found 84% use or plan to use AI tools, Google's DORA research found 90% use AI, and JetBrains found 85% use it regularly. Daily use is lower, at 50.6% of professional developers in 2025.

Which AI coding agent do developers use most?

In Stack Overflow's 2026 survey, Claude Code was the most-used coding agent at 66% of respondents, followed by GitHub Copilot at 59%. In 2025 the most-used AI tools overall were ChatGPT at 81.7% and GitHub Copilot at 67.9%. Respondents are self-selected, so the figures are not market share.

Do developers trust AI-generated code?

Mostly with conditions. In 2025, 46% of Stack Overflow respondents distrusted AI output and 33% trusted it, and 66% said answers are almost right but not quite. In 2026, 48% said they trust AI when they can easily validate the answer. Google's DORA survey found only 24% trust it a lot.

Does AI actually make developers faster?

Self-reports say yes, and a controlled trial says it depends. Over 80% of DORA respondents reported higher productivity, but METR's randomized trial found experienced open-source developers took 19% longer with early-2025 AI tools, while believing they were 20% faster. The trial had 16 developers, so it does not generalize to every team.

Related