Usage ranking
AI coding agents by usage on April 1, 2026
This ranks how much each agent is used and followed, not how good it is: a popular agent can be a worse fit for your work than a small one. Click an agent to see where its numbers come from. Each source's column shows the agent's scaled score (the leader is 100) over the raw number. 3 sources had data on this day. The score averages an agent's scaled values over the sources that can measure it (at least three).
1
Claude CodeAnthropic · closed sourceUsage score 4710011M2897k138.8M- npm 100 · 11M
- GitHub 28 · 97k
- VS Code 13 · 8.8M
- npm: 10011,325,282 downloads of @anthropic-ai/claude-code in the 7 days to Apr 1, 2026. That's the most of any agent in this source, so it scores 100. Source
- GitHub: 2896,883 stars on github.com/anthropics/claude-code on Apr 1, 2026. The leader here is OpenClaw with 344k, so Claude Code scores 97k ÷ 344k × 100 = 28. Source
- VS Code: 138,753,902 installs of the anthropic.claude-code extension on Apr 1, 2026. The leader here is GitHub Copilot with 69M, so Claude Code scores 8.8M ÷ 69M × 100 = 13. Source
Usage score: (100 + 28 + 13) ÷ 3 = 47. Claude Code website
2
OpenClawOpenClaw · open sourceUsage score 38151.7M100344k–- npm 15 · 1.7M
- GitHub 100 · 344k
- npm: 151,681,155 downloads of openclaw in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so OpenClaw scores 1.7M ÷ 11M × 100 = 15. Source
- GitHub: 100343,810 stars on github.com/openclaw/openclaw on Apr 1, 2026. That's the most of any agent in this source, so it scores 100. Source
Not counted: VS Code, because it has no VS Code extension.
Usage score: (15 + 100) ÷ 3 = 38, divided by 3 rather than 2 because every agent's average counts at least 3 sources. OpenClaw website
3
GitHub CopilotGitHub · closed sourceUsage score 354444k–10069M- npm 4 · 444k
- VS Code 100 · 69M
- npm: 4443,739 downloads of @github/copilot in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so GitHub Copilot scores 444k ÷ 11M × 100 = 4. Source
- VS Code: 10069,165,482 installs of the GitHub.copilot-chat extension on Apr 1, 2026. That's the most of any agent in this source, so it scores 100. Source
Not counted: GitHub, because it has no public GitHub repo.
Usage score: (4 + 100) ÷ 3 = 35, divided by 3 rather than 2 because every agent's average counts at least 3 sources. GitHub Copilot website
4
Codex CLIOpenAI · open sourceUsage score 20323.6M2171k96.4M- npm 32 · 3.6M
- GitHub 21 · 71k
- VS Code 9 · 6.4M
- npm: 323,568,948 downloads of @openai/codex in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Codex CLI scores 3.6M ÷ 11M × 100 = 32. Source
- GitHub: 2170,609 stars on github.com/openai/codex on Apr 1, 2026. The leader here is OpenClaw with 344k, so Codex CLI scores 71k ÷ 344k × 100 = 21. Source
- VS Code: 96,441,680 installs of the openai.chatgpt extension on Apr 1, 2026. The leader here is GitHub Copilot with 69M, so Codex CLI scores 6.4M ÷ 69M × 100 = 9. Source
Usage score: (32 + 21 + 9) ÷ 3 = 20. Codex CLI website
5
OpenCodeAnomaly · open sourceUsage score 156726k39134k–- npm 6 · 726k
- GitHub 39 · 134k
- npm: 6726,152 downloads of opencode-ai in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so OpenCode scores 726k ÷ 11M × 100 = 6. Source
- GitHub: 39134,191 stars on github.com/anomalyco/opencode on Apr 1, 2026. The leader here is OpenClaw with 344k, so OpenCode scores 134k ÷ 344k × 100 = 39. Source
Not counted: VS Code, because there's no snapshot of its extension page from around that day.
Usage score: (6 + 39) ÷ 3 = 15, divided by 3 rather than 2 because every agent's average counts at least 3 sources. OpenCode website
6
Gemini CLIGoogle · open sourceUsage score 127769k29100k1584k- npm 7 · 769k
- GitHub 29 · 100k
- VS Code 1 · 584k
- npm: 7769,350 downloads of @google/gemini-cli in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Gemini CLI scores 769k ÷ 11M × 100 = 7. Source
- GitHub: 2999,785 stars on github.com/google-gemini/gemini-cli on Apr 1, 2026. The leader here is OpenClaw with 344k, so Gemini CLI scores 100k ÷ 344k × 100 = 29. Source
- VS Code: 1583,958 installs of the Google.gemini-cli-vscode-ide-companion extension on Feb 17, 2026. The leader here is GitHub Copilot with 69M, so Gemini CLI scores 584k ÷ 69M × 100 = 1. Source
Usage score: (7 + 29 + 1) ÷ 3 = 12. Gemini CLI website
7
ClineCline · open sourceUsage score 8182k1760k53.3M- npm 1 · 82k
- GitHub 17 · 60k
- VS Code 5 · 3.3M
- npm: 182,315 downloads of cline in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Cline scores 82k ÷ 11M × 100 = 1. Source
- GitHub: 1759,755 stars on github.com/cline/cline on Apr 1, 2026. The leader here is OpenClaw with 344k, so Cline scores 60k ÷ 344k × 100 = 17. Source
- VS Code: 53,325,272 installs of the saoudrizwan.claude-dev extension on Mar 16, 2026. The leader here is GitHub Copilot with 69M, so Cline scores 3.3M ÷ 69M × 100 = 5. Source
Usage score: (1 + 17 + 5) ÷ 3 = 8. Cline website
8
Roo CodeRoo Code · open sourceUsage score 3–723k21.4M- GitHub 7 · 23k
- VS Code 2 · 1.4M
- GitHub: 722,506 stars on github.com/RooCodeInc/Roo-Code on Mar 5, 2026. The leader here is OpenClaw with 344k, so Roo Code scores 23k ÷ 344k × 100 = 7. Source
- VS Code: 21,426,494 installs of the RooVeterinaryInc.roo-cline extension on Apr 1, 2026. The leader here is GitHub Copilot with 69M, so Roo Code scores 1.4M ÷ 69M × 100 = 2. Source
Not counted: npm, because it has no command-line package on npm.
Usage score: (7 + 2) ÷ 3 = 3, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Roo Code website
9
Qwen CodeAlibaba · open sourceUsage score 2196k621k–- npm 1 · 96k
- GitHub 6 · 21k
- npm: 195,725 downloads of @qwen-code/qwen-code in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Qwen Code scores 96k ÷ 11M × 100 = 1. Source
- GitHub: 621,485 stars on github.com/QwenLM/qwen-code on Apr 1, 2026. The leader here is OpenClaw with 344k, so Qwen Code scores 21k ÷ 344k × 100 = 6. Source
Not counted: VS Code, because it has no VS Code extension.
Usage score: (1 + 6) ÷ 3 = 2, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Qwen Code website
10
Kilo CodeKilo · open sourceUsage score 2035k517k1849k- npm 0 · 35k
- GitHub 5 · 17k
- VS Code 1 · 849k
- npm: 035,282 downloads of @kilocode/cli in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Kilo Code scores 35k ÷ 11M × 100 = 0. Source
- GitHub: 517,441 stars on github.com/Kilo-Org/kilocode on Apr 1, 2026. The leader here is OpenClaw with 344k, so Kilo Code scores 17k ÷ 344k × 100 = 5. Source
- VS Code: 1849,303 installs of the kilocode.Kilo-Code extension on Mar 15, 2026. The leader here is GitHub Copilot with 69M, so Kilo Code scores 849k ÷ 69M × 100 = 1. Source
Usage score: (0 + 5 + 1) ÷ 3 = 2. Kilo Code website
11
CrushCharm · open sourceUsage score 202.5k621k–- npm 0 · 2.5k
- GitHub 6 · 21k
- npm: 02,543 downloads of @charmland/crush in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Crush scores 2.5k ÷ 11M × 100 = 0. Source
- GitHub: 620,724 stars on github.com/charmbracelet/crush on Mar 2, 2026. The leader here is OpenClaw with 344k, so Crush scores 21k ÷ 344k × 100 = 6. Source
Not counted: VS Code, because it has no VS Code extension.
Usage score: (0 + 6) ÷ 3 = 2, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Crush website
12
piMario Zechner · open sourceUsage score 207.6k619k–- npm 0 · 7.6k
- GitHub 6 · 19k
- npm: 07,588 downloads of @mariozechner/pi-coding-agent in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so pi scores 7.6k ÷ 11M × 100 = 0. Source
- GitHub: 619,019 stars on github.com/badlogic/pi-mono on Mar 3, 2026. The leader here is OpenClaw with 344k, so pi scores 19k ÷ 344k × 100 = 6. Source
Not counted: VS Code, because it has no VS Code extension.
Usage score: (0 + 6) ÷ 3 = 2, divided by 3 rather than 2 because every agent's average counts at least 3 sources. pi website
13
AuggieAugment Code · closed sourceUsage score 0034k–1700k- npm 0 · 34k
- VS Code 1 · 700k
- npm: 033,761 downloads of @augmentcode/auggie in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Auggie scores 34k ÷ 11M × 100 = 0. Source
- VS Code: 1699,841 installs of the augment.vscode-augment extension on Feb 9, 2026. The leader here is GitHub Copilot with 69M, so Auggie scores 700k ÷ 69M × 100 = 1. Source
Not counted: GitHub, because it has no public GitHub repo.
Usage score: (0 + 1) ÷ 3 = 0, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Auggie website
14
Letta CodeLetta · open sourceUsage score 007.8k11.7k–- npm 0 · 7.8k
- GitHub 1 · 1.7k
- npm: 07,755 downloads of @letta-ai/letta-code in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Letta Code scores 7.8k ÷ 11M × 100 = 0. Source
- GitHub: 11,740 stars on github.com/letta-ai/letta-code on Mar 4, 2026. The leader here is OpenClaw with 344k, so Letta Code scores 1.7k ÷ 344k × 100 = 1. Source
Not counted: VS Code, because it has no VS Code extension.
Usage score: (0 + 1) ÷ 3 = 0, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Letta Code website
15
AmpSourcegraph · closed sourceUsage score 0037k–099k- npm 0 · 37k
- VS Code 0 · 99k
- npm: 036,744 downloads of @sourcegraph/amp in the 7 days to Apr 1, 2026. The leader here is Claude Code with 11M, so Amp scores 37k ÷ 11M × 100 = 0. Source
- VS Code: 098,538 installs of the sourcegraph.amp extension on Feb 19, 2026. The leader here is GitHub Copilot with 69M, so Amp scores 99k ÷ 69M × 100 = 0. Source
Not counted: GitHub, because it has no public GitHub repo.
Usage score: (0 + 0) ÷ 3 = 0, divided by 3 rather than 2 because every agent's average counts at least 3 sources. Amp website
Not ranked that day, only one source has them: OpenHands (GitHub, 68k), Aider (GitHub, 41k), Hermes Agent (GitHub, 21k), Droid (npm, 1.2k).
Method
How the usage score is made
No single public number measures how much each AI coding agent is used, so we blend the ones that exist. In each source the leader scores 100 and the rest are scaled to it. An agent's score is the average over the sources that can measure it. Sources that can't see an agent at all (Cursor has no npm package) are left out of its average, but the average always counts at least three sources, so an agent seen by only one or two can't outrank one that leads everywhere. An agent needs two sources to be ranked.
- npmCLI downloads, 7 days · data from Apr 1, 2026
Downloads of each agent's command-line package in the 7 days to the date, with one-off bursts (mirrors, bots) capped. Downloads count installs and CI runs, not people, and agents with their own installer are undercounted. Source
- GitHubStars · data from Apr 1, 2026
Stars on the agent's GitHub repo, from monthly snapshots. Stars measure interest, and build up over time. Source
- VS CodeExtension installs, all time · data from Apr 1, 2026
All-time installs of the agent's VS Code extension, from monthly snapshots. Cursor and Windsurf install from Open VSX instead, which isn't counted. Source
Browse
The most used AI coding agent, month by month
Pick the model behind your agent
Most of these agents let you choose the model. Send one prompt to up to six models and see which answer you'd actually use, with what each one cost.
Data through October 21, 2026 for 20 agents. npm downloads from the npm registry; OpenRouter app rankings from OpenRouter (CC BY 4.0); GitHub stars and VS Code installs from monthly snapshots in the Internet Archive and the live APIs; survey figures as published by Stack Overflow and JetBrains.