Amplifying/agent-intelligence
/coding-agentsclaude-code

Anthropic

Claude Code

Terminal-first coding agent for repository editing, shell commands, code review, and long-running implementation workflows.

CLIIDEDesktopWebCloudMobileActiveProprietaryBYOKMCP Yes

12.5M

Attributed output

GitHub public-attribution count, 2026-06-09

#1

Index rank

of 34 tracked agents

$2.5B+

Business signal

Claude Code run-rate revenue

$20/mo Pro

Entry price

Dynamic shared 5-hour usage window

Not published

GitHub stars

No public repo

9,440,524

npm demand/wk

Adjusted from raw npm channel

342,693,416

npm installs to date

Since 2025-02-24

Activity Over Time

Weekly attributed output on public GitHub. Output visible through the agent's own bot or workflow is exact; output measured through declared attribution is an estimate. Window: 2025-02-03 to 2026-06-08.

Model mixtrailing 8 weeks, 2026-04-06 to 2026-06-14

Which Claude model its users actually run, from visible public attribution metadata on Claude-marked commits.

Opus 66.2%Sonnet 32.9%Haiku 0.9%

npm Installs Over Time

Weekly npm downloads since 2025-02-24, the install-demand trajectory.

Market

Developer output

12.5M

Public GitHub commits attributed to Claude

Claude Code declares attribution on the work it produces, so coverage is strong. The headline is a point-in-time estimate that moves between measurements and a floor: squash-merges, private repositories, and a separate review integration sit outside it.

Last full week (2026-06-08): 3.7M signed commits (-11.6% week over week), search estimate.

Distinct users: 6,756 humans opened the 12,600 most recently sampled PRs, about 1.9 PRs per person.

Merge rate: 91.5% of its pull requests get merged (2,077,567 of 2,270,518 over 8 settled weeks, 2026-04-13 to 2026-06-07). The window ends about ten days back so recently-opened PRs have had time to merge.

GitHub review activity: 4.7% of its merged PRs had at least one GitHub review record before merging; the other 95.3% had none. This is a broad weekly signal and can include bot reviews; the human-review read below comes from the enriched PR sample.

In significant repos (company-owned or 10+ stars), 33.3% of its merged PRs got a human review. Across all sampled repos, the rate is 24.2%. Human review means a non-author, non-bot GitHub reviewer in the PR-lifecycle sample.

Median time to merge: 25 min. The unreviewed median is 0.19h. The human-reviewed median is 3.97h. From a 120-PR sample merged 2026-05-07 to 2026-06-05.

Test-inclusion: 42.5% of its sampled merged PRs touch a test file. This is a rough read on production readiness. 3 large PRs hit GitHub's 100-file path read cap, so treat this as a floor.

Flow share: 13.2% of the 7998 most recent public PRs sampled (1057 PRs, 8 windows, last 2026-06-17).

Task lens: 4.2M agent-opened PRs. The declared mark can be removed by the team, so treat this as approximate.

Snapshot 2026-06-09, confidence Medium, tracked since 2026-06-09, +25.6% over 6 days. Read the methodology.

Index rank
#1 of 34
Demand share
36.5% of tracked npm demand floor
GitHub share
No public repo
npm channel
11.8M raw downloads/week (Native preferred)

Best fit

Terminal-first engineeringLarge refactorsClaude-native teamsAgentic code review
  • Enterprise use represents over half of Claude Code revenue, according to Anthropic.
  • Anthropic says Claude Code weekly active users doubled after January 1, 2026, but it does not publish the absolute WAU count.
  • Claude Code docs now recommend native install and package-manager routes, so npm downloads are treated as one visible install channel rather than total usage.
  • Anthropic says business subscriptions to Claude Code quadrupled since the start of 2026.
  • Anthropic cites SemiAnalysis estimating Claude Code authors 4% of all public GitHub commits, double the prior month.
  • Anthropic's May 2026 Series H valued the parent company at $965B and said company run-rate revenue crossed $47B earlier that month.
  • By February 2026 Claude Code's run-rate had reached about $2.5B, roughly half a billion above Cursor's at the time, and developers reported switching from Cursor to Claude Code (Bloomberg, via Business Insider).
  • Anthropic was, at one point, heavily dependent on Cursor: a Cursor employee estimated it was 40-50% of Anthropic's revenue before Claude Code launched (Business Insider).
npm channel caveat: Native preferredExpand

Claude Code's npm downloads halved in late April 2026 — a distribution change, not a usage decline.

In late April 2026 Anthropic moved the recommended install to native installers (Homebrew, WinGet, apt, dnf, apk). The npm package still works and installs the same binary, so the drop from ~13M to ~8M weekly is installs shifting off npm, not fewer users. npm now counts only one channel and understates total adoption.

Demand floor
9.4M / wk

Trailing 4-week npm average. Native installs are unobserved, so this remains a floor.

Latest npm week
11.8M

Complete Mon-Sun week when cron-backed

Peak npm week
13.1M

Observed public npm range

4-week average
9.4M

As of 2026-06-14

All sourced signals (7)Expand
SignalValueTypeConfidence
npm weekly downloads9,738,911Public metricHigh
Recommended install pathNative installer; npm remains supportedOfficialHigh
Claude Code run-rate revenue$2.5B+ and more than doubled since Jan. 1, 2026OfficialHigh
Claude Code business subscriptionsQuadrupled since Jan. 1, 2026OfficialHigh
Claude Code enterprise mixEnterprise use is over half of Claude Code revenueOfficialHigh
Public GitHub commit shareSemiAnalysis estimated 4% of public GitHub commits were authored by Claude CodeReportedMedium
Anthropic valuation$965B post-money in May 2026 Series HOfficialHigh

Where Claude Code Works

Languages
TypeScript38%
Python22%
JavaScript7%
Rust7%
Go5%
HTML5%
Frameworks
React25%
Next.js20%
Flutter8%
FastAPI7%
Astro6%
SwiftUI4%
Industries
AI/ML43%
Devtools12%
Finance5%
Robotics4%
Gaming4%
Media4%

Share of Claude Code's own classified sampled work in each segment. Full breakdowns: Languages, Frameworks, Industries.

Pricing And Economics

Expand

Claude Pro is $20/month; Max is $100/month for 5x Pro or $200/month for 20x Pro.

Entry price
$20/mo Pro
Top self-serve price
$200/mo
Pricing unit
Dynamic shared 5-hour usage window
Token transparency
Retail token equivalent
Included usage
Published average guidance: Pro ~10-40 Claude Code prompts/5h; Max 5x ~50-200; Max 20x ~200-800. Actual capacity varies with prompt length, context size, model, file attachments, tool use, and demand.
Overage path
After plan limits, users can wait for reset or enable extra usage / Anthropic Console API credits at standard API rates.

Retail Token Equivalent

No fixed subscription token quota is published. At current Sonnet API retail, $100 equals about 33.3M input or 6.7M output tokens; $200 equals about 66.7M input or 13.3M output tokens. This is dollar-equivalent math, not a Max plan entitlement.

PlanPriceAllowanceUnitToken / Overage Notes
Claude Pro$20/mo~45 Claude messages or ~10-40 Claude Code prompts every 5 hours, with actual usage varying by context and model.5-hour prompts$20 at Sonnet API retail = ~6.7M input tokens or ~1.3M output tokens. Manual API-credit path after limit; otherwise wait for reset. Claude Code on Pro is Sonnet-only in published guidance.
Claude Max 5x$100/mo~225 Claude messages or ~50-200 Claude Code prompts every 5 hours, with actual usage varying by context and model.5-hour prompts$100 at Sonnet API retail = ~33.3M input / ~6.7M output; at Opus retail = ~20M input / ~4M output. Can switch to API credits or wait for reset. Claude Code can switch from Opus to Sonnet at 20% of Max 5x usage.
Claude Max 20x$200/mo~900 Claude messages or ~200-800 Claude Code prompts every 5 hours, with actual usage varying by context and model.5-hour prompts$200 at Sonnet API retail = ~66.7M input / ~13.3M output; at Opus retail = ~40M input / ~8M output. Can switch to API credits or wait for reset. Claude Code can switch from Opus to Sonnet at 50% of Max 20x usage.

Unit Economics (Measured)

Mean API-retail cost per completed task on the Artificial Analysis coding agent index (358 tasks across 3 suites, pass@1 over 3 runs, snapshot 2026-06-09). These are pay-per-token API costs, not subscription effective costs.

HarnessAA indexCost / taskTokens / taskTime / task
Claude Code, Opus 4.8 (max)77.2$4.624.8M18m
Claude Code, Opus 4.8 (medium)67.2$1.601.7M9m
Claude Code, Opus 4.7 (max)66.6$4.145.7M14m
Claude Code, Opus 4.7 (medium)59.9$1.241.7M6m
Claude Code, Qwen3.7 Plus (thinking)53.1$4.983.3M11m
Claude Code, GLM-5.152.7$2.264.8M22m
Claude Code, Opus 4.6 (medium)51.3$1.272.2M7m
Claude Code, Kimi K2.650.5$0.763.7M42m
Claude Code, DeepSeek V4 Pro (high)50.1$0.353.5M18m
Claude Code, Sonnet 4.6 (medium)49.4$1.022.2M9m

Break-even vs API retail on the top harness (Claude Code, Opus 4.8 (max), $4.62/task): Claude Pro at $20/mo covers about 4 tasks; Claude Max 5x at $100/mo covers about 22 tasks; Claude Max 20x at $200/mo covers about 43 tasks. Assumes the Artificial Analysis task mix.

Included with Claude Pro and Max plans, or billable through Anthropic API credits. Pro is light-use; Max 5x is $100/month; Max 20x is $200/month.

  • Anthropic estimates Pro users get roughly 10-40 Claude Code prompts every 5 hours.
  • Max 5x is roughly 50-200 Claude Code prompts every 5 hours.
  • Max 20x is roughly 200-800 Claude Code prompts every 5 hours.
  • Opus consumes limits faster than Sonnet; Max can automatically switch from Opus to Sonnet at usage thresholds.

Company And Funding

Expand
Ownership
Private
Founded
2021
Headquarters
San Francisco, CA
Employees
~5,200
Total raised
~$132B
Most recent round
Series H, $65B (Apr 2026)
Round leads
Altimeter Capital, Dragoneer, Greenoaks, Sequoia Capital, Coatue, GIC

Parent-company total. Claude Code is one of several Anthropic products, not a separately financed entity.

Notable investors

GoogleAmazonSequoia CapitalLightspeedSalesforce VenturesMenlo Ventures

Enterprise And Security

Expand
SSO / SCIM
Yes
Deployment
VPC / dedicated
Trains on your code
Not trained on your code
Zero data retention
Enterprise tier

Compliance

SOC 2 Type IIISO 27001HIPAA (BAA available)GDPR

Admin controls

SAML SSO / SCIMAudit logsRole-based access

Anthropic does not train models on code sent under commercial terms. Enterprise customers can add a zero-data-retention addendum, and Claude Code can route through Amazon Bedrock or Google Vertex in a customer cloud account.

Product

Providers

Anthropic

Model families

Claude SonnetClaude OpusClaude Haiku

Capabilities

Repo editingTerminal commandsPlanningSubagentsMCPCode reviewWeb search

Tradeoffs

  • Claude-only by default
  • Plan limits are usage-based estimates rather than fixed message counts