I've been using Claude Code since its launch in May 2025. Since then, the tool has become the most-used AI coding agent on the market. In 8 months, Claude Code overtook GitHub Copilot and Cursor.
In this article, you'll find the latest numbers, data and facts about Claude Code. From GitHub stars and downloads to market share and benchmarks to pricing and technical specifications.
- Claude Code is the most-used AI coding tool (Pragmatic Engineer Survey, 906 developers, Feb. 2026) with 46% 'most loved' rating
- 148,179 GitHub stars, 533M+ cumulative npm downloads, estimated annualized revenue of $2.5 billion (Anthropic, February 2026)
- 58% on SWE-bench Verified (standalone), 95% first-try correctness, React dashboard built in 47 minutes
- Since September 22, 2026, Claude Opus 5.5 is the default model in Claude Code, scoring 66.4% on Terminal-Bench 4.0 at roughly 40% lower cost than Opus 5
1. What Is Claude Code?
Claude Code is Anthropic's AI coding agent for the terminal. You describe what you want to build in natural language. Claude Code writes the code, creates files, runs tests and commits changes.
The tool understands entire codebases. It reads files, analyzes dependencies and makes decisions autonomously. No copy-pasting between browser and editor. Everything happens directly in the terminal.
Here are the key milestones:
From beta to market leader in 8 months. No other AI coding tool has made that leap this fast. For a detailed walkthrough, check out our Claude Code guide.
2. GitHub Stars and Downloads
Claude Code's source is publicly viewable on GitHub, but the tool isn't officially open source: Anthropic licenses Claude Code under its own commercial terms. The numbers still show strong growth since launch:
- 148,179 GitHub stars (September 26, 2026, GitHub API)
- 533M+ cumulative npm downloads (since the February 2025 launch through September 25, 2026, npm API)
- 57M npm downloads in the last 30 days (August 26 through September 24, 2026, npm API)
- Current version: 2.1.283
- Update frequency: Multiple releases per week
For comparison: GitHub Copilot had significantly fewer GitHub stars after 8 months. That shows how much excitement Claude Code has generated in the developer community.
Claude Code's reach extends well beyond its own stars and downloads. According to a SemiAnalysis analysis from February 2026, Claude Code already accounted for roughly 4% of public GitHub commits at that point, having doubled within a single month. At that pace, SemiAnalysis projects Claude Code will account for more than 20% of all daily commits by the end of 2026.
3. User Numbers and Adoption
The Pragmatic Engineer Survey from February 2026 (906 developers surveyed) provides the most comprehensive data on AI coding tool adoption to date:
- 73% of engineering teams use AI coding tools daily (up from 41% in 2025)
- 95% of developers use AI tools at least weekly
- 75% use AI for more than half of their coding work
- 70% use 2 to 4 AI tools simultaneously
The bottom line: AI coding tools are no longer a niche. They're part of the daily workflow. And Claude Code is at the top.
The distribution by company size is particularly interesting. Adoption depends heavily on company size:
- Small companies: Claude Code dominates with 75% adoption rate
- Enterprise (10,000+ employees): GitHub Copilot leads with 56%, Claude Code falls behind
The pattern is clear. Small, agile teams prefer Claude Code. Large enterprises lean toward GitHub Copilot. The reason: enterprise customers often have existing GitHub contracts and stricter compliance requirements.
That said: 71% of developers who regularly use AI agents use Claude Code. When it comes to agentic coding (the future of programming), Claude Code already dominates.
4. Market Share in AI Coding
Market share depends heavily on the use case. For complex tasks, Claude Code leads by a wide margin according to estimates:
Complex tasks include multi-file refactoring, architecture decisions, large-scale debugging and building entire features. Exactly what a terminal agent like Claude Code was built for.
For simple autocomplete, the picture looks different. According to estimates, GitHub Copilot still leads here:
GitHub Copilot has its strength in quick autocomplete directly in the editor. Claude Code, on the other hand, excels at tasks that require context across multiple files. For a detailed comparison, check out our article on Claude Code vs. OpenAI Codex.
A much larger survey shows the same shift even more clearly. The JetBrains Developer Ecosystem Survey 2026 polled more than 15,000 professional developers worldwide between May and July 2026, with a direct comparison to the same year's January survey:
Source: JetBrains Developer Ecosystem Survey 2026 (15,000+ developers, May-Jul. 2026).
Claude Code more than doubled its workplace usage between January and summer 2026, and in this survey it now sits well ahead of GitHub Copilot. GitHub Copilot, meanwhile, is losing share, while OpenAI Codex grew in parallel from 3% to 16%. Across all tools, the same survey found that 90% of respondents use AI coding agents at least weekly, and 68% daily.
5. Popularity and Satisfaction
In the same survey, developers were asked which AI coding tool they value most. The results speak for themselves:
46% of developers name Claude Code as their "most loved" tool. That's more than double Cursor (19%) and five times GitHub Copilot (9%).
The problem: GitHub Copilot was first to market. Many developers use it out of habit or because their company provides it. When it comes to actual satisfaction, Claude Code is far ahead.
The high popularity has concrete reasons. Claude Code understands the context of entire projects, works autonomously and delivers consistently usable results. More on this in our Claude Code commands reference.
6. Revenue and Economic Impact
Claude Code is not just popular. It's also a massive revenue driver for Anthropic:
- Estimated annualized revenue (run rate) in February 2026: approximately $2.5 billion
- $1 billion run rate reached in just 6 months after launch
- Anthropic total revenue in May 2026: approximately $47 billion ARR (per Anthropic Series H announcement)
From 0 to $1 billion in 6 months. No other AI coding tool has reached this mark this fast. For comparison: Cursor crossed the $500 million mark after more than a year on the market.
For context: the total AI coding market is estimated at roughly $15 to $20 billion in 2026. Claude Code alone already accounts for a significant share of that.
Anthropic as a whole has grown well past the $47 billion ARR reported in May. According to Bloomberg and TechCrunch, the company's total annualized run rate had already crossed $65 billion by the end of July 2026. Anthropic hasn't broken out Claude Code's specific share of that figure since February.
7. Coding Benchmarks
Popularity numbers are one thing. Objective benchmark results are another. Claude Code performs strongly in both categories:
58% on SWE-bench Verified may sound less impressive than the 88.6% achieved by Claude Opus 4.8. Claude Opus 5 replaced it as Anthropic's coding recommendation in July, before Fable 5.1 became Anthropic's new API frontier model on September 1. Anthropic reported 55.8% on Terminal-Bench 4.0 for Fable 5.1 and 60.9% for Mythos 5.1. Since September 22, Claude Opus 5.5 has been available and, as of Claude Code version 2.1.280, is the default model on the Pro, Max, Team, and Enterprise plans, replacing Sonnet 5. Anthropic reports 66.4% on Terminal-Bench 4.0 for Opus 5.5 (Fable 5.1: 55.8%, Opus 5: 52.3%). Claude Sonnet 5.5, the sonnet alias in Claude Code since September 28, reaches 70.6%. Anthropic estimates that translates to roughly 40% lower operating cost than Opus 5 on typical workloads. The plain list prices per million tokens are $4 instead of $5 for input and $20 instead of $25 for output, a 20% reduction. These vendor figures use a different setup than SWE-bench Verified. They establish neither a SWE-bench score nor that every Claude Code plan automatically uses the newest model. Claude Code as a tool additionally uses context, file access, and iterative debugging.
The 95% first-try correctness is more relevant in practice. You give Claude Code a task, and in 19 out of 20 cases the result works immediately. No revisions needed.
8. Pricing and Costs
Claude Code is included in all Claude subscriptions. As of version 2.1.280, Opus 5.5 is the default model on Pro, Max, Team, and Enterprise; Pro previously defaulted to Sonnet 5. Here's the overview:
My recommendation: start with Claude Pro for $20/month. That's enough for most developers. If you work with Claude Code for hours every day, the upgrade to Max is worth it.
8.1. Price Comparison with Competitors
Price-wise, Claude Code is on par with Cursor. GitHub Copilot is cheaper but offers fewer agentic features. The question is ultimately what you need: simple autocomplete or a full-fledged coding agent.
9. Technical Specifications
The technical specs of Claude Code at a glance:
The context window deserves special attention. 200,000 tokens is roughly 150,000 words or about 500 pages of text. Claude Code can understand entire codebases with hundreds of files at once.
With 1 million tokens of context, Claude Code is among the tools with the largest context window. Since version 2.1.277, Claude Code automatically reads an AGENTS.md file if a project has no CLAUDE.md. Learn more about configuring projects with CLAUDE.md files in our guide.
10. Competitor Overview
The AI coding market is fiercely competitive. Here are the key players in a direct comparison:
Claude Code's biggest advantage is its combination of a terminal-based workflow and strong Claude models. Copilot has the larger ecosystem. Cursor offers the best IDE integration. The Antigravity CLI scores with Gemini 3.1 Pro's 1M context window, but heavily throttles its free tier.
For a much more detailed comparison, check out our article on Claude Code vs. OpenAI Codex.
11. Interesting Facts and Records
To wrap up, here are the most noteworthy facts about Claude Code:
- Fastest overtake: Claude Code launched in May 2025 and overtook GitHub Copilot (on the market since June 2021) and Cursor as the most loved AI coding tool within just 8 months.
- Terminal philosophy: While most AI coding tools started as IDE extensions, Anthropic deliberately chose the terminal. The reason: maximum flexibility and full control over file access.
- CLAUDE.md as a standard: The concept of project configuration files (CLAUDE.md) caught on so well that competitors introduced similar formats (GEMINI.md, .cursorrules).
- $1 billion in 6 months: Claude Code reached the $1 billion run rate faster than any other AI coding tool in history.
- 95% first-try correctness: 19 out of 20 code outputs work on the first attempt. That's among the highest reported rates among AI coding agents.
- Not open source: Despite common assumptions, Claude Code isn't an open-source tool. Anthropic licenses Claude Code under its own commercial terms. The code on GitHub is publicly viewable, but it isn't freely usable or modifiable.
- 47 minutes for a dashboard: In a standardized benchmark, Claude Code built a complete React dashboard (with authentication, data visualization and API integration) in under an hour.
- 75% adoption at startups: Three out of four startups use Claude Code. No other AI coding tool has an adoption rate this high among small companies.
The bottom line: Claude Code has shaken up the AI coding market in record time. The combination of a strong model, terminal-based workflow and rapid release cadence has clearly struck a nerve. I'm curious to see how the numbers develop over the coming months.
If you want to try Claude Code yourself, our Claude Code guide has everything you need to get started. We've also put together an overview of the most important Claude Code commands.
