AI Agent Statistics 2026: Live Usage, Growth and Activity Data
Current statistics on AI agents from public data: package downloads, editor installs, GitHub stars, pull requests opened by coding agents, release pace and benchmark results. Updated daily.
Every figure on this page is computed from public sources and refreshed daily. Numbers are as of Nov 20, 2026 and cover the 492 agents, frameworks and agent platforms tracked on AgentGid. You are welcome to quote them with a link to this page; the methodology explains how each is collected.
Headline numbers
- 319M package downloads of tracked agents and frameworks in the last 7 days (npm and PyPI combined), up 4.8% on the previous week.
- 158M cumulative installs of tracked agent extensions on the Visual Studio Marketplace.
- 10.8M GitHub stars across tracked repositories.
- 545 stable releases shipped by tracked projects in the last 7 days.
- 1.2M pull requests opened by 7 coding agents on public GitHub repositories last week, of which 1.1M have been merged so far.
Download counts include automated installs such as CI runs and auto-updates, so they measure activity, not people.
Most downloaded agents and frameworks
| # | Project | Category | Downloads, 7 days | Change vs previous 7 days | Change over 30 days |
|---|---|---|---|---|---|
| 1 | AI SDK | Frameworks | 47.2M | +3.9% | +38.4% |
| 2 | LangChain | Frameworks | 41.1M | +5.7% | -2.7% |
| 3 | Claude Agent SDK | Frameworks | 29.3M | +5.9% | +37.6% |
| 4 | OpenAI Codex | Terminal | 26.6M | +2.4% | +7.4% |
| 5 | LiteLLM | Agent tooling | 24.4M | +5.0% | +4.6% |
| 6 | LangGraph | Frameworks | 20.1M | +5.5% | +29.2% |
| 7 | Claude Code | Terminal | 18.9M | +15.1% | +15.5% |
| 8 | Langfuse | Agent tooling | 14.7M | -0.5% | +33.5% |
| 9 | Strands Agents | Frameworks | 9M | +1.3% | +3.8% |
| 10 | OpenAI Agents SDK | Frameworks | 7.2M | +4.8% | +38.9% |
Most installed agent extensions in VS Code
| # | Extension | Installs | Average rating |
|---|---|---|---|
| 1 | GitHub Copilot | 79.4M | 2.7 / 5 (423 ratings) |
| 2 | Claude Code | 30.6M | 3.7 / 5 (936 ratings) |
| 3 | OpenAI Codex | 16.9M | 2.7 / 5 (975 ratings) |
| 4 | Tabnine | 9.7M | 4.0 / 5 (623 ratings) |
| 5 | Cline | 6M | 3.9 / 5 (344 ratings) |
| 6 | CodeGPT | 2.6M | 3.4 / 5 (337 ratings) |
| 7 | Amazon Q Developer | 1.9M | 3.9 / 5 (59 ratings) |
| 8 | Kilo Code | 1.7M | 3.8 / 5 (246 ratings) |
Installs are cumulative since each extension was published.
Most starred repositories
| # | Project | Category | GitHub stars |
|---|---|---|---|
| 1 | OpenClaw | General | 395K |
| 2 | ECC | Agent tooling | 307K |
| 3 | DeepSeek Harness | General | 284K |
| 4 | Hermes Agent | General | 265K |
| 5 | OpenCode | Terminal | 228K |
| 6 | n8n | Automation | 211K |
| 7 | Firecrawl | Agent tooling | 203K |
| 8 | AutoGPT | General | 188K |
| 9 | Dify | Automation | 161K |
| 10 | Langflow | Automation | 157K |
Fastest growing by downloads
Projects with at least 20,000 downloads in the last 30 days, ranked by growth against the 30 days before.
| # | Project | Downloads, 30 days | Growth |
|---|---|---|---|
| 1 | Novita Sandbox | 252K | +284% |
| 2 | AgentScope | 434K | +246% |
| 3 | AI | 496K | +163% |
| 4 | Inkeep | 683K | +145% |
| 5 | agentmemory | 89.5K | +133% |
| 6 | Momentic | 1.7M | +118% |
| 7 | Metaharness | 400K | +117% |
| 8 | PR-Agent | 98.2K | +104% |
More on the movers page.
Pull requests opened by coding agents
Some coding agents can be identified on public GitHub by a bot account or branch prefix. Counts for the week of Nov 9, 2026:
| Agent | Pull requests opened | Merged so far | Merge rate, previous 4 weeks |
|---|---|---|---|
| Claude Code | 798,602 | 753,528 | 95% |
| OpenAI Codex | 277,056 | 237,131 | 87% |
| Cursor | 119,155 | 95,311 | 80% |
| Devin | 10,773 | 7,102 | 67% |
| GitHub Copilot | 9,125 | 5,849 | 70% |
| Jules | 942 | 495 | 46% |
| OpenCode | 676 | 531 | 82% |
This covers public repositories only, and only agents that leave a recognisable signature. Details and weekly charts are on the GitHub activity page.
Benchmark results
On Terminal-Bench 4.1, which scores complete agents on long terminal tasks:
| Agent | Best run | Tasks solved | Cost of the run |
|---|---|---|---|
| Claude Code | Opus 5.5 | 65.8% ± 3.1 | $4,719 |
| OpenAI Codex | GPT-6.2 Astra | 62.1% ± 2.3 | $1,409 |
| GitHub Copilot | Copilot Fusion | 71.2% ± 2.9 | $1,987 |
| mini-SWE-agent | Inkling Small | 91.0% ± 94.0 | $50 |
| Antigravity CLI | Gemini 4 Carbon | 64.5% ± 3.2 | $1,358 |
| Grok Build | Grok 4.7 | 38.8% ± 3.6 | $3,683 |
Source: Terminal-Bench (tbench.ai). All boards are on the benchmarks page.
How the field is split
| Category | Tracked | Open source | With a free option |
|---|---|---|---|
| AI coding agents | 87 | 40 | 65 |
| AI app builders | 9 | 0 | 8 |
| AI agents for game development | 13 | 1 | 10 |
| AI design agents | 8 | 2 | 8 |
| Agent frameworks and SDKs | 51 | 49 | 50 |
| Agent tooling for cost, context and memory | 65 | 59 | 60 |
| Agent builders and workflow automation | 17 | 10 | 13 |
| General-purpose AI agents | 44 | 32 | 42 |
| Browser and computer-use agents | 26 | 17 | 24 |
| AI research agents | 14 | 8 | 14 |
| AI data analysis agents | 9 | 3 | 6 |
| AI agents for business teams | 76 | 9 | 26 |
| AI agents for DevOps and security | 17 | 7 | 8 |
| Agent orchestration and team workspaces | 39 | 31 | 35 |
| AI video and media agents | 15 | 12 | 15 |
| AI trading and investing agents | 2 | 2 | 2 |
Pricing
Among the 45 coding agents with a listed plan for individuals, entry prices run from $0 (a free tier) to $500 a month, and the middle one costs $20.
In customer support, 16 of the 34 tracked agents publish a per-unit price (per resolution, conversation or credit) instead of, or on top of, a monthly plan.
About these statistics
Everything here comes from public registries and APIs: npm, PyPI, GitHub, the Visual Studio Marketplace and others listed on the data sources page. Nothing is estimated or supplied by vendors. That also sets the limits: closed products that are not distributed through public channels are under-represented, and none of these figures is a count of users or revenue.
Agents mentioned in this guide
| # | Agent | Gid Score | Downloads 7d | 7d | 30d | Stars | Latest release | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 |
Claude CodeAnthropic |
88 | 18.9M | up 15.1% | up 15.5% | 156K+136/day | 2.1.328yesterday | ||||
| 2 |
OpenAI CodexOpenAI |
84 | 26.6M | up 2.4% | up 7.4% | 136K+144/day | 0.168.07d ago | ||||
| 1 |
LangChainLangChain |
79 | 41.1M | up 5.7% | down 2.7% | 150K+76/day | 1.5.22d ago | ||||
| 4 |
n8nn8n |
79 | 107K | down 28.5% | up 4.9% | 211K+167/day | 1.123.982d ago | ||||
| 5 |
|
75 | — | — | — | — | — | ||||
| 8 |
OpenClawOpenClaw Foundation (nonprofit); created by Peter Steinberger |
71 | 3.1M | up 5.8% | down 24.3% | 395K+120/day | 2026.10.332d ago | ||||
Live table, updated daily (figures as of Nov 20, 2026); numbers in the text above are as of the dates given there. "#" is the rank among products of the same kind by Gid Score, our 0-100 composite of adoption, community, attention and momentum (how it is calculated); a dash means no public data.
Sorted by rank. Promo marks an official promotion the maker runs right now (a discount, free access or credits); open it for the details on Deals.