iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Sometimes—but the answer depends on the task and on what “faster” means. A controlled GitHub Copilot experiment found developers completed one defined JavaScript task 55% faster. In a different randomized study, experienced developers took 19% longer to complete real issues in familiar, mature open-source repositories when AI tools were allowed. A UK public-sector trial reported time savings, but those figures came from participant surveys rather than a randomized measurement of hours saved. These results describe different settings and measures; none establishes one productivity gain for all developers.
What the studies measured—and what they found
The findings below are not directly interchangeable. Some studies timed task completion; others asked participants to estimate time saved or report their experience. The tools, participants and coding contexts also differed.
| Study | People and coding context | Measure | Finding and qualification |
|---|---|---|---|
| GitHub Copilot controlled experiment, reported by GitHub in 2022 and updated in 2024; summarized by Microsoft Research in 2023 | 95 professional developers implemented a JavaScript HTTP server. | Timed task completion and completion rate. | The Copilot group averaged 1 hour 11 minutes, compared with 2 hours 41 minutes without Copilot: GitHub reported a 55% speed gain, with a 95% confidence interval of 21% to 89% and P=.0017. Completion rates were 78% with Copilot and 70% without. Microsoft Research reported a 55.8% faster completion time for the same underlying experiment, not a separate replication. The result applies to this defined task, not automatically to larger or longer-running software work. |
| METR randomized trial, July 2025 | 16 experienced open-source developers completed 246 real issues—bugs, features and refactors—in mature repositories they had worked in for years. The repositories averaged more than 22,000 stars and one million lines of code. Participants primarily used Cursor Pro with Claude 3.5 or 3.7 Sonnet, as well as other tools they chose. | Measured task completion time, alongside participants’ estimates of AI’s effect. | Tasks took 19% longer when AI was allowed. Before the study, participants predicted 24% faster completion; afterward, they estimated AI had made them 20% faster. The measured and perceived effects diverged. METR says the result is a snapshot of early-2025 tools in this setting, not a verdict on most developers, other domains or later tools. |
| UK Government Digital Service trial, November 2024–February 2025 | 2,500 licenses were distributed across more than 50 public-sector organizations. The main survey analysis included 424 responses from 31 departments; 73% of respondents reported at least five years of coding experience. | Participant-reported time savings and experience, plus Copilot telemetry. | Respondents reported an average of 56 minutes saved per working day, including 24 minutes on code creation or analysis. Sixty-five percent said they completed tasks faster, 67% reported less time searching for examples or information, and 56% reported more efficient problem solving. These are survey estimates, not a randomized control-group estimate of actual time saved. The report notes possible overlap among task estimates and optimism that could inflate savings. |
| METR follow-up, update published February 24, 2026 | A follow-up study begun in August 2025 included returning and newly recruited developers, but METR identified selection and time-measurement problems. | Raw speed estimates with confidence intervals; METR assessed whether they reliably represented productivity. | The raw estimates were an 18% speedup for returning participants (interval: 38% speedup to 9% slowdown) and a 4% speedup for new participants (interval: 15% speedup to 9% slowdown). Both intervals include no effect. METR says the data are unreliable as a measure of real productivity impact, so these numbers should not be treated as a current speedup estimate. |
Why “developer productivity” is not one number
Task completion time is one useful outcome, but it does not capture every way a coding assistant might help or hinder. GitHub frames productivity through the SPACE framework, which includes satisfaction and well-being, performance, activity, communication and collaboration, and efficiency and flow. A developer can feel more focused or satisfied without completing a particular task faster; likewise, faster code generation alone does not establish that a team shipped more valuable, correct software.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallIn GitHub’s accompanying survey of more than 2,000 technical-preview users—primarily professional developers (about 60%), with students (about 30%) and hobbyists (about 7%) also represented—73% said Copilot helped them stay in flow and 87% said it helped preserve mental effort during repetitive tasks. Those are reports of perceived experience, not measured completion-time gains. The same distinction matters in the METR trial: participants’ estimates of a speedup did not match the measured slowdown.
#1 Best Overall
- 🎙️ Hands-Free Voice Typing for Windows & Mac – Powered by iOS & Android dictation technology, AI VoiceWriter allows fast, accurate speech-to-text directly on your desktop. Simply speak, and your words appear in real time. Compatible with Windows 10 & above, macOS 13 & above.
- ✍️ AI Writing Assistant for Effortless Editing – Boost productivity with AI proofreading, rephrasing, and formatting. Perfect for emails, reports, creative writing, and professional content.
- 💻 Works Seamlessly in Any Desktop App – Type with your voice in Microsoft Word, Google Docs, PowerPoint, Teams, emails, and more. Just place your cursor in any text field and start speaking!
- 📱 Mobile App for Enhanced Voice Input – The AI VoiceWriter mobile app enhances voice recognition by using your phone’s microphone as an input device for clearer, more accurate dictation—while typing on your desktop. Supports iOS 15 & above, Android 9.0 & above.
- 🌎 Multilingual Voice Typing & AI Assistance – Supports 33 languages for dictation, plus AI-powered features in Chinese, English, Japanese, Korean, French, German, Spanish, Italian and, Swedish.
Code acceptance is another separate measure. In the UK trial, GitHub Copilot telemetry showed an average code-line acceptance rate of 15.8%, and 39% of users said they had committed AI-suggested code. Acceptance does not by itself show that code was correct, valuable, or faster to produce; nor does a low acceptance figure alone establish that a tool had no benefit.
How to interpret a productivity claim
Before applying a study’s result to your own work, check what was actually compared. These questions help distinguish a narrow benchmark result from evidence about day-to-day development:
- What kind of task was timed? A self-contained exercise and a bug fix or refactor in a large, mature codebase place different demands on a developer.
- Did developers know the repository? Familiarity can affect how quickly someone can locate relevant code, understand conventions and judge a proposed change.
- Who participated? Experience, role and selection into a study can shape the result; a small group of experienced repository contributors is not a sample of every developer.
- Which tools and when? Identify the assistant, model and study dates. The METR trial tested early-2025 tools, while its 2026 update did not yield a reliable estimate of the effect of newer tools.
- Was time measured or estimated? Randomized task timing, survey responses and retrospective estimates answer different questions.
- What counted as a successful result? Look for completion rates, quality checks, review or testing criteria, not just how quickly code was produced.
- Is the claim about an individual or a team? Saving time on one task does not by itself establish higher team throughput, improved collaboration or better software outcomes.
What the evidence supports
The strongest conclusion is conditional: AI coding tools can speed up some work, but their effect varies with the task, codebase, developer and measurement method. The Copilot experiment demonstrates a substantial gain on a bounded JavaScript task; METR’s 2025 trial found a slowdown in a demanding, familiar-repository setting; and the public-sector figures describe reported experience rather than controlled time savings. METR’s 2026 follow-up does not resolve the disagreement because the organization says its estimate is unreliable.
That is why the headline percentages should not be averaged into a universal speedup. For a developer or team deciding whether an assistant helps, the relevant test is whether it improves the outcomes they care about—such as time to a tested, accepted change, quality, focus or collaboration—in their own work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
