MyTechDailyNews Guide to Octoverse 2026: AI Agents, Model Menus, and Builders From Outside Tech
Most technology coverage reaches readers through a product launch or a funding announcement. This one arrives as an activity ledger: Octoverse, GitHub's annual measurement of how software gets made, records code pushed 17.9 billion times across the past 12 months, and the more interesting story hides inside that total — who, and what, is doing the pushing. For anyone who depends on an app, a hospital system, a payment service or a public website, the mechanics of how code is produced shape what appears on their screen next year.
The report's own framing is narrower than its consequences. Read for a general tech audience, three shifts stand out: automated assistants now open a large share of the work, no single AI model dominates the toolbox, and the people building software increasingly come from fields that were never counted as tech at all.
Five numbers to hold on to
- The volume is unprecedented. Code was pushed 17.9 billion times in the past 12 months, a one-year total larger than 2024 and 2025 combined, while public contributions alone reached 2.5 billion in 2026.
- Assistants write the first drafts now. By September, more than 40% of pull requests began with an agent; in March that share was 17%. Among repositories created this year, the average one carrying an agent configuration file logged eight times as many commits as an average new repository without one.
- Choice is spreading, not consolidating. In September, 61% of GitHub Copilot users who selected a model themselves used more than one, and paid plans saw 29 distinct models in May and 47 in September, with not a single model carrying more than 25% of requests that month.
- Builders arrived from new fields. For the first time public repositories are viewed by industry, and repositories serving mining, construction, health care and agriculture have each grown several times over since 2022.
- The forecasts aged badly. One year into last year's five-year developer growth projections, every country modelled has already closed at least a fifth of its distance to 2030, and the forecasting models no longer agree on where things go next.
Concrete projects behind an abstract total
Scale is easier to picture through specific builds. In Nepal, one developer published a flood bulletin on GitHub Pages carrying verified information on roads, rescue, weather and power; it later matured into Gati, a Nepali-first relay of official disaster alerts running on GitHub Pages and Actions. At NASA's Jet Propulsion Laboratory, engineers are producing software designed to run programs aboard spacecraft. Elsewhere, enthusiasts built open-source firmware that lets people customise their own e-readers. Curiosity pointing in every direction is what generated the headline figure.
The pace: commits, automations, and a mid-year break upward
Activity changed hands at a rate the platform had not registered before. Commits across public and private repository types climbed to 34.3 billion, more than three times where they stood two years ago, and merged pull requests more than tripled to 1.3 billion over the same period, averaging 3.5 million a day.
Automation carried much of the load. GitHub Actions executed a job 1,440 times every second this year — the equivalent of 124 million times a day, and more than three times last year's total. Maintenance ran on its own conveyor too: developers merged more than 30 million Dependabot pull requests to keep dependencies current and retire known vulnerabilities, about double 2025's pace.
Telemetry traces the shape of the surge. Activity climbed steadily from January, then bent sharply upward around June, as if the whole field spotted the same opening simultaneously. Right then a cluster of industry-wide shifts landed: stronger underlying models, better harnesses and tooling built around them, and agents that changed character — no longer single-session helpers but persistent workers able to track progress across days, orchestrate other specialised agents, and lean heavily on the cloud, executing remotely as well as on local machines.
Two crowds drive the momentum. From June onwards, about 55% of all commits came from brand-new repositories. Plenty more came from long-time builders: developers who pushed code in the past 12 months have been on GitHub for an average of 36 months, and of those who opened a pull request in August, more than 75% pushed code again in September. Their footprint is widening as well — the average developer now works across 10% more repositories than a year ago.
Open source shows the same staying power. Public contributions soared to 2.5 billion in 2026, more than double last year's total of 1.12 billion. Individual work gives those numbers a face: @kinlane recorded more than 1 million contributions across his public repositories, using agentic workflows to keep a machine-readable catalogue of public API information current; @kim-em logged more than 178,000 contributions while maintaining mathlib, a community library of computer-checked mathematics; and @andig, a maintainer of the open source EV charging controller and home energy management system evcc, contributed over 21,000 times.
More developers, from more places than the old map showed
Every statistic above belongs to a person. GitHub now serves more than 225 million of them, and the community is thickening around the globe. The United States (38.3 million) and India (32 million) remain the two largest developer communities, each adding more than 10 million people this year. Brazil added 3.28 million developers, the third most of any country.
| Country | Total number of developers |
|---|---|
| United States | 38.3 M |
| India | 32.0 M |
| China | 12.4 M |
| Brazil | 10.2 M |
| Japan | 7.15 M |
| United Kingdom | 6.47 M |
| Indonesia | 6.34 M |
| Germany | 6.30 M |
| Russia | 5.07 M |
| France | 4.75 M |
Last year the projection was that India would reach 57.5 million developers by 2030 and the United States 54.7 million. One year into that five-year window, every country in the projections has already closed at least a fifth of the distance to its own 2030 figure: 38.6% for the United States, 28.3% for India.
Regional detail supplies the local context behind that growth:
| Region | Growth driver | Market context |
|---|---|---|
| Americas | Fintech and digital payments |
|
| EMEA | AI adoption in manufacturing and startups, plus a fast-growing developer community in Africa |
|
| APAC | Government investment in AI infrastructure and talent |
|
With this much growth absorbed in twelve months, the next five years are harder to call. More than 40 forecasting models were run forward to 2031 — time-series and regression models built on historical sign-up and usage data, each assuming a different pace. About half call for extreme growth, with developer counts on GitHub more than tripling by 2031; just under half call for substantial growth, doubling to tripling; only three models expect anything close to modest, under 2x growth. Forecasts remain estimates, and a spread this wide measures how much the 2026 surge widened what is plausible. What holds steady is the shape of the trend: a developer community growing in more places at once.
Industries the report had never been measuring
In earlier editions of this report, everywhere meant more developers and more countries, which is still visible in the ranking of the top 10 countries by pushes to public repositories, adjusted by profile location.
| Rank | Country |
|---|---|
| 1 | United States |
| 2 | India |
| 3 | China |
| 4 | Brazil |
| 5 | Germany |
| 6 | United Kingdom |
| 7 | Korea |
| 8 | France |
| 9 | Japan |
| 10 | Canada |
This year the picture gained a second axis: more fields, including industries that had not been measured before. Public repositories are sorted with the North American Industry Classification System, or NAICS, which groups economic activity into sectors. A language model classified random samples of public repositories created each year from 2022 to 2026, assigning each to the industry it serves from its name, description, topics and README. Two signals were tracked: whether a repository mentions AI, and whether an agent opened a pull request against it.
Those signals do not always move together. In administrative and support services, mining, health care and agriculture, more than four in 10 repositories mention AI, yet agent-opened pull requests are rare — less than 3 percent. Construction breaks the pattern: about a quarter of its repositories mention AI, while about 9 percent have agent-opened pull requests, the highest share of any sector, with manufacturing next.
Re-weighted by growth since 2022, sectors well beyond software are climbing. Professional, scientific, and technical services rose from fourth to second place, behind only information. Agriculture, forestry, fishing and hunting climbed two spots, and finance and insurance and health care and social assistance each moved up one. Against each sector's own 2022 baseline, mining, quarrying, and oil and gas extraction grew roughly 65 times from a small starting base, administrative and support services and construction 17 to 18 times, and even manufacturing, the slowest of the group, almost tripled.
Behind the sector shifts are people who know their own fields and now build with AI coding agents. In manufacturing, a small team created a way to turn a photo of your tools into a custom-fit, 3D-printable drawer organizer. A health care research lab made it easier for scientists to analyse huge digital scans of tissue samples, the kind pathologists study to understand disease. In Copenhagen, a university lab built a pipeline that classifies video of chickens at play, from frolicking to worm chasing. A University of Waikato energy centre produced a toolkit that helps industrial plants find heat they can reuse, and Bulgaria's digital ministry open-sourced a portal where anyone can trace who won a public contract and for how much. For writers, a novelist released a free app that makes a manuscript look like a book as you type. In field after field, the people closest to a problem are building the solution, with agents helping along the way.
Professional, scientific, and technical services — up about 5 times over its own 2022 baseline and now second place — is the work of researching, engineering and advising, and open code has long been part of how it is done. The work behind AlphaFold earned Demis Hassabis and John Jumper a share of the 2024 Nobel Prize in Chemistry. Tensor2Tensor accompanied the paper that introduced the Transformer, the architecture behind today's large language models, Attention Is All You Need. sars-cov-2-stability holds the code and data behind one of the earliest papers on how long the virus survives in the air and on surfaces. eht-imaging was one of two open-source libraries used to reconstruct the first image of a black hole. A repository has quietly become part of the paper trail behind published research.
A workflow flip: describe the job, then review what comes back
More developers now begin a pull request by specifying the outcome they want and orchestrating the agents that carry it out. By September 2026, more than 40% of pull requests on GitHub started with an agent's first draft, up from 17% in March — 110 million in September alone, about 6x March's volume.
Agent-started changes also reach further. About 87% of them alter multiple files, compared with 79% of pull requests without detected agent involvement. They land quickly too: in February, 83% of agent-involved pull requests merged within two weeks, a touch behind the roughly 86% for requests without detected agent involvement. By September, agent pull requests were at 90% while the comparison group held at about 86%.
Delegation is deepening. Among GitHub Copilot's monthly active users, 55% now use agentic features, up from 35% a year ago, and the steepest growth is in the most hands-off mode: the Copilot cloud agent, where a developer assigns a task and an agent works on it independently, grew 230% year over year, against 85% for agent chat. Developers still ask agents questions, but increasingly they hand over whole tasks — public repositories mentioning interoperability or reverse engineering increased more than 300% year over year, from about 30,000 created in 2025 to more than 130,000 in 2026.
Review is where developers shape what comes back, and more tooling is being built for it. Stacked pull requests, which split related work into smaller changes that are easier to review in sequence, launched as a preview on July 30, and more than 264,000 developers had used them by September. Usage of AI-assisted code review and GitHub Advanced Security both grew more than 90% year over year. The platform also helps secure that work: push protection, which screens for exposed secrets such as API keys and passwords, ran 5.8 billion scans this year, nearly four times 2025's pace.
A model shelf that keeps growing, and nobody settling on one item
Developers match models to the work they hand to their agents. In September 2026, 61% of GitHub Copilot users who selected a model themselves used more than one model across their sessions, up from 50% four months earlier. The lineup expanded at the same time: developers on paid GitHub Copilot plans used 29 distinct models in May and 47 in September, and no single model carried more than 25% of requests that month. Some decisions come down to cost and familiarity, with price and functionality weighed against need and a lean toward providers already trusted.
To see how choices spread, models used in GitHub Copilot were grouped by what they are built for — deep reasoning and planning, general purpose, and fast and lightweight — and requests were counted only where a developer explicitly selected a model. That excludes cases where GitHub Copilot picks through auto model selection, GitHub and Microsoft staff usage, background utility models, duplicate requests used for monitoring and testing, and requests made with a developer's own API key. Out of every 100 developer-selected requests, 38 went to general purpose models, 37 to deep reasoning and planning models, and 25 to fast and lightweight models, so about 25% of requests go to a fast model.
Viewed as a time series over the past three months, tracking request share alongside how many developers tried each model and how many kept returning, three patterns stand out:
- New releases are used heavily before most developers have tried them, so early adopters lean on a model hard before the rest follow.
- A model stays in rotation after usage cools: when its share of requests falls, the share of developers still using it falls more slowly.
- Developers blend kinds of models, spreading requests across all three groups, with few staying with any one kind for long.
That spread points to a lot of choosing: a lineup that keeps growing, a best fit that shifts from task to task, and all three kinds of models held in rotation. Many developers also let the tool decide: in September 2026, 63% of paying GitHub Copilot developers used auto model selection at least once. Planning, implementing and checking a change each place different demands on a model in capability, speed and cost, which is why Project HydraFusion began as a research preview of model orchestration, matching models to each step; in the CLI and the GitHub Copilot app it routed almost 60,000 sessions. That work is now generally available as GitHub Copilot Fusion, bringing model orchestration together with auto model selection to choose the right model behind the scenes.
Programming languages gain a new voice in the choice
Language choice has long reflected what developers want to build, what their teams know, and what existing systems require. This year's data raises a fourth possibility: the language an agent picks when a developer does not specify one.
The order set last year held in 2026 — TypeScript first, Python second, JavaScript third — but the pace picked up. From September 2025 to September 2026, TypeScript added 3.62 million monthly contributors, Python 3.23 million, and JavaScript 1.82 million, the largest gains among the top 10 languages. Rust grew fastest in percentage terms among the top languages, doubling over the year. Extending recent trends, forecasts put TypeScript and Python near 13 million monthly contributors each by September 2027, JavaScript at 7.4 million, and every other language in the top 10 below about 2.2 million.
The fastest-growing language communities are also where AI appears most, charted by setting each language's yearly growth against the share of its merged pull requests with detected AI involvement. The steepest growers sit in a band of about 40% to 44% AI involvement, a second group lands across about 21% and 34%, and more than one in five of the remaining leading languages hold near 24%.
Four misconceptions worth retiring
| Common assumption | What the report actually shows |
|---|---|
| Agents are replacing developers. | Agents make the first move, but review is still where developers shape the result: more than 40% of pull requests start with an agent, and by September agent pull requests merged at 90% versus about 86% for non-agent work. |
| Everyone has settled on one best model. | 61% of users who choose their own model use more than one, and not a single model carried more than 25% of requests in September as the shelf went from 29 distinct models in May to 47. |
| This is still purely a software-industry story. | Repositories serving mining, construction, health care and agriculture each grew several times since 2022, and construction now shows the highest share of agent-opened pull requests at about 9 percent. |
| Growth is concentrated in the usual hubs. | Brazil added 3.28 million developers, Africa is described as a fast-growing developer community, and within one year every projected country had closed at least a fifth of its distance to 2030. |
FAQ: the questions this report keeps prompting
What is Octoverse, in one line?
It is GitHub's annual read-out of how software actually gets built, drawn from platform activity rather than opinion surveys — which is why the 17.9 billion pushes are its headline.
Why should a non-technical reader care?
Because the same toolchain produced a Nepali disaster-alert relay, a public-contract tracing portal, and an app that renders a novel as a book while it is written. Changes in how software is made arrive eventually as changes in the products people use.
Does this mean software simply gets cheaper to make?
The report measures activity, not pricing. What it does show is output per person moving: agent-started pull requests touch multiple files in about 87% of cases, and 83% merged within two weeks early in the year, rising to 90% by September.
Is one AI model clearly winning?
No. Requests split 38 general purpose, 37 deep reasoning and planning, and 25 fast and lightweight out of every 100 developer-selected requests, and 63% of paying Copilot developers tried auto model selection at least once rather than deciding themselves.
How reliable are the 2031 predictions?
Treat them as a range. More than 40 forecasting models were run forward to 2031: roughly half expect extreme growth, more than tripling; just under half expect substantial growth, doubling to tripling; only three models land under 2x growth.
Five practical steps readers can take now
- Treat agent output as a first draft. Concentrate your review effort on multi-file changes, since about 87% of agent-started pull requests alter more than one file.
- Match the model to the step instead of the habit. With 38 of every 100 selected requests going to general purpose, 37 to deep reasoning and planning, and 25 to fast and lightweight, a single default is already out of date.
- Automate upkeep before you automate features. More than 30 million Dependabot pull requests were merged to keep dependencies current and reduce known vulnerabilities, about double 2025's pace.
- Screen for secrets on the way out. push protection ran 5.8 billion scans this year, nearly four times 2025's pace, a reminder that exposed API keys and passwords are a volume problem, not an anecdote.
- Watch outside your own sector. Professional, scientific, and technical services moved to second place, mining grew roughly 65 times from its 2022 base, and the closest people to a problem are increasingly the ones building its solution.
Strip away the arithmetic and a simpler conclusion remains: the people writing the software that runs daily life are more numerous, more widely distributed, and more heavily assisted than any five-year plan assumed, and they are working in industries that would once have seemed far from this world. For readers of MyTechDailyNews, the practical lesson is to watch what gets built in health care, agriculture, construction and government as closely as the consumer gadget cycle — that is where the next familiar product will come from. And to the millions of developers contributing to work like the examples above: the effort behind the things you use is visible, and it is worth paying attention to.
Coverage compiled for MyTechDailyNews by Sarah Chen, Tech Daily newsroom. Questions or corrections about this story: hello@techdaily.com.