AI Coding Agent trends in May 2026

We have been operating a site called AI Coding.Info since July 2025.

This is a site that observes usage trends related to AI Coding Agents such as Claude Code, Codex CLI, Github Copilot, and Gemini from a fixed point from information in Github repositories. To determine the use of AI Coding Agent, we conduct daily surveys under the following conditions.

Daily inspection of 9,000 public repositories on Github
- TOP 300 Github stars by programming language
- Researched from 30 types of programming languages
Targets 16 types of AI Coding Agents
Only when the rule files used by various AI Coding Agents are in the Github repository, it is determined that AI Coding Agent is being used.

Last month's trends

https://ai-coding.info/reports/articles/20260504

AI Coding Agent usage rate is 11.1%

AI Coding Agent repository utilization rate was 11.1%, an increase of 1.1% from 10.0% last time.

This time, I would like to particularly focus on the usage rate of TypeScript's AI Coding Agent.

**TypeScript's AI Coding Agent usage rate has exceeded 50%. **

https://ai-coding.info/languages/typescript?date=2026-06-01#adoption-rate

I wonder if this day will come someday. I thought so, but it seems like it has finally arrived.

We are now in a time when people who do not use AI Coding Agent in TypeScript are in the minority

It seems that has arrived. This in itself is quite shocking. I think so.

AI Coding Agent usage rate on June 1, 2026

https://ai-coding.info/?date=2026-06-01#adoption-rate

AI Coding Agent usage rate on May 1, 2026

https://ai-coding.info/?date=2026-05-01#adoption-rate

Trends in AI Coding Agent usage rate from May 1, 2026 to June 1, 2026

https://ai-coding.info/?date=2026-06-01&since=2026-05-01&until=2026-06-01#share-trend

Share of AI Coding Agent by product

The share by product is as follows.

Ranking	Product name	Share rate
1st place	Codex CLI	36.9%
2nd place	Claude Code	32.8%
3rd place	Copilot Agent	16.8%
4th place	Gemini CLI	6.1%
5th place	Cursor	5.1%

Personally, I think "Claud Code's share will decrease and Codex CLI's share will increase." That was my prediction, but surprisingly it wasn't. Regarding Codex CLI, there was a large increase in share by 1.8%, but regarding Claude Code, the share was maintained at the same level as last month. There was a surprising aspect. On the other hand, I feel that Github Copilot's share is decreasing a little. Recently, there has been a revision of Github Copilot's price plan (https://github.blog/jp/2026-04-28-github-copilot-is-moving-to-usage-based-billing/) and there has been some controversy, so the downward trend may continue. In May, there was Google I/O, and there were announcements surrounding Antigravity 2.0. Last month, Gemini had a 5.8% share, and this month it is 6.1%, so it has increased slightly. This is the tendency. Regarding Antigravity, we have not included any logic to measure the share, but it may have had a positive impact on the existing Gemini to some extent.

AI Coding Agent share rate on June 1, 2026

https://ai-coding.info/?date=2026-06-01&since=2026-05-01&until=2026-06-01#agent-share

AI Coding Agent share rate on 5/1/2026

https://ai-coding.info/?date=2026-05-01&since=2026-05-01&until=2026-06-01#agent-share

AI Coding Agent usage status by programming language

The programming language in which AI Coding Agent is used the most is "TypeScript," the second is "Rust," the third is "Python," the fourth is "Go," and the fifth is "C#." TypeScript remains the most used language, but Rust, Python, and Go have become more stable. That's the impression I get. This area is up and down between the two armies. That's what it feels like. "C#/Kotlin/C++" is one of a kind. It's hard to say, but I guess this is the line for the third army. That's the impression. Should you include or not include "Javascript/Ruby/Java" below? There are places like that, and I have a strong impression of 1st army, 2nd army, and others.

AI Coding Agent ranking by programming language on June 1, 2026

https://ai-coding.info/?date=2026-06-01&since=2026-05-01&until=2026-06-01#language-agent-rank

AI Coding Agent ranking by programming language on May 1, 2026

https://ai-coding.info/?date=2026-05-01&since=2026-05-01&until=2026-06-01#language-agent-rank

Monthly trends in the number of repositories used by AI Coding Agent

The number of repositories was 1,436 as of May 1, 2026, but it has increased to 1,611 as of June 1, 2026, and the number of repositories using AI Coding Agent has increased by 175.

Changes in the number of repositories using AI Coding Agent from 2026/5/1 to 2026/6/1

https://ai-coding.info/?date=2026-06-01&since=2026-05-01&until=2026-06-01#time-based-bar-chart

Will OSS that doesn't use AI disappear?

As mentioned earlier, according to the data on June 1, 2026, TypeScript's AI Coding Agent usage rate was 53%, more than half. Among programming languages, AI Coding has by far the highest penetration rate. So,

**How much have GitHub's popular repositories changed compared to the past? Is that influence due to AI? **

I would like to take a look at that. This time, we will compare the data of GitHub's TOP 300 repositories surveyed on September 1, 2025 with the data of GitHub's TOP 300 repositories surveyed on June 1, 2026.

First of all, we crawl the TOP300 every day, but there may be some errors or surpluses, so there may not be exactly 300. The number of repositories checked on September 1, 2025 was 312, and the number of repositories checked on June 1, 2026 was 299.

How many repositories from September 1, 2025 remain as of June 1, 2026?

In the survey on September 1, 2025, 312 repositories were investigated, and at that time, 21.5% (67 repositories) were confirmed to be using AI, and 78.5% (245 repositories) were not using AI. As of September 1, 2026, there are 234 cases. therefore,

**25% of repositories fell outside of the TOP300 in 9 months. **

Now, if we think about repositories that have fallen outside of the TOP300,

Of the 245 cases where AI was not used, 61 cases, or 24.9%, were outside the top 300.
Out of 67 cases of AI usage, 17 cases, or 27.9%, fell outside the top 300.

Therefore, in terms of "falling outside the TOP 300", there seems to be no significant difference between using AI and not using AI.

What is the structure of the remaining repositories as of June 1, 2026?

First, what happened to the 67 repositories that used AI as of September 1, 2025? Based on the discussion earlier, the total number has been reduced to 50 as of June 1, 2026.

48 out of 50 cases where AI was used, 96% continued to use AI.
2 out of 50 cases where AI was used, 4%, stopped using AI.

That's the conclusion. Personally, I'm surprised that there are two cases, but overall

**Once AI is adopted, it continues to be used. **

That's basically what I understand. So, what about repositories that do not use AI? Similarly, the population is reduced to 184. Thinking from that,

120 of the 184 repositories without AI, or 62.2%, remained without AI.
Of the 184 repositories not using AI, 64 repositories (34.8%) were converted to using AI.

From this point on, approximately 1/3 of the repositories that were not using AI and maintained the TOP300 changed their policy to using AI. ** You can see that.

What is the structure of the repositories that have newly entered the TOP300 as of June 1, 2026?

From the graph, there are 65 new repositories that are ranked in the TOP300.

Of the 65 repositories newly added to the TOP300, 16 (24.6%) do not use AI.
Of the 65 repositories newly added to the TOP300, 49 (75.4%) use AI.

It becomes. **Approximately 3/4 of the newly added repositories to the TOP300 are powered by AI. **

Think in terms of inflow and outflow.

I would like to analyze the TOP300 ranking from the perspective of inflows and outflows. The necessary facts are arranged as follows.

Of the 245 cases where AI was not used, 61 cases, or 24.9%, were outside the top 300.
Out of 67 cases of AI usage, 17 cases, or 27.9%, fell outside the top 300.
Of the 65 repositories newly added to the TOP300, 16 (24.6%) do not use AI.
Of the 65 repositories newly added to the TOP300, 49 (75.4%) use AI.

From here,

49 AI-based repositories were added to the TOP300, and 16 were leaked. This is a total increase of 33 cases in 9 months.
16 repositories that were not used by AI entered the TOP300, and 61 repositories were leaked. This is a total decrease of 45 cases in 9 months.

In the TOP300, the number of **AI-based repositories is steadily increasing. ** That can be said.

Personal hypothesis and conclusion

“In the nine-month period from September 2025 to June 2026, the usage rate of AI Coding Agent in TypeScript’s top 300 stars skyrocketed from 21.5% to 53.8%.”

From that,

"OSS that does not use AI has largely disappeared, and the configuration repositories in the top 300 rankings may have changed significantly."

That was the hypothesis. However, the reality was not like that. That is the conclusion. First, looking at repositories that fell outside of the TOP 300, the percentage of repositories that did not use AI withdrew was 24.9%, and the percentage of repositories that used AI withdrew was 27.9%. It can be said that they are almost the same, and the withdrawal rate is higher when using AI. It can also be taken as Now, discussing the remaining repositories, 1/3 of the repositories that originally did not use AI converted to use AI, and 2/3 remained unused. Repositories that originally used AI will continue to use AI for the most part. That's what it means. Finally, among the repositories newly ranked in the TOP300 repositories, 3/4 use AI and 1/4 do not use AI. That's what it means. As a story,

"All existing OSS rankings have been wiped out by AI"

Although this is an interesting result, the reality is that this is not the case.

**Repositories that were not using AI have been converted to using AI and remain at the top of the rankings. **

That's what I mean. In a way, if you think about it, I don't think there was any need to analyze it. That's what it felt like. When you look at The AI Coding Agent usage rate has increased sharply from 21.5% to 53.8%,'' you may have intuitively understood that 47.2% who are not using AI are existing OSS, and there are few new entrants who are not using AI.'' However, the question Are the repositories using AI Coding Agent new or existing?'' cannot be answered with just the prerequisite data, which I think we found out through this analysis. I had a feeling that Once the OSS becomes large, it's difficult to change the system, so it might be difficult to introduce AI,'' but I think it's amazing that about 1/3 of existing OSS has switched to using AI in 9 months.'' Personally, I thought, I haven't written any code by hand for over half a year now.'' However, since I'm in the software development industry, I'm used to this sense of speed, and if I were working in a different industry, I guess it wouldn't be possible. I think so. In that sense, the rate of penetration of AI may be unusually fast considering other industries. On the other hand, the wave of AI will engulf everything. There is a feeling that many of the newly ranked repositories use AI. In this analysis, repositories that do not use AI and repositories that use AI have a similar probability of falling outside the rankings. As a result, the rate of AI usage in the overall rankings will naturally continue to increase. In addition, 1/3 of OSS that did not use AI have converted to using AI. Considering that,

"Will OSS that doesn't use AI disappear?"

The answer to the question is

"The proportion of OSS using AI among the TOP 300 repositories will increase"

That's what it means. This is not a question of Can a new OSS be a giant killer of an existing OSS if AI is used?'' but rather, Existing OSS is evolving by using AI in the same way.'' In other words, we found that AI is not weeding out existing OSS, but rather that AI is penetrating both existing OSS and new OSS. AI is not a weapon only for existing powers, nor is it only for emerging powers.

"AI is equal in a sense"

It may be.

Will OSS that does not use AI disappear? ~Summary of AI Coding trends in May 2026 as seen through data~