Skip to content
Sign in
↑Part of: Anthropic reports $16.3 billion in first-half sales as IPO filing details steep losses
Technology · North America

Anthropic says Claude produces most of its code

Anthropic says Claude now produces more than 90% of the code contributions at the company. Engineers’ average daily code contributions rose to more than 45 times their 2024 level, while Claude also supports research, development and safety work on future models.

AI-assisted synthesis · Based on 5 outlets' reporting · Article text read from 1 of 5 outlets · 1 article captured only in partHow it's made

View timeline · 9 entries
5 outlets2 regionsLatest reporting 2 min readNo new reporting in 3 days
Coverage: Center OnlyL 0 · C 4 · R 0 · 1 unrated
Updat3 graphic: Anthropic
Updat3 graphic

What happened

Key facts

  • Anthropic says Claude produced more than 90% of the company’s code contributions.
  • Average daily code contributions per engineer in the July-to-September quarter were more than 45 times the average daily level in 2024.
  • Anthropic says it uses its models to research, design and develop later models, a process it calls closing the loop.
  • Anthropic says Claude has helped improve algorithms to make better use of its computing resources.

The brief

Anthropic says its Claude models produced more than 90% of contributions to the company’s codebase in September, according to The Times of India. The company also reported that engineers’ average daily code contributions in the July-to-September quarter were more than 45 times the average daily level in 2024.

The figures describe the company’s reported use of its own AI in software development. Anthropic said it uses its models to research, design and develop the models that follow them. The company calls this process “closing the loop.” It also said Claude helped improve algorithms so the company could make better use of its computing resources.

The Times of India reported that Anthropic’s workforce increased from about 940 employees at the end of 2024 to roughly 5,300 by late September. Anthropic said its strategy is to continue growing the business without adding staff at the same pace. The account provides the workforce figures alongside the company’s description of its approach to expansion.

The company reported regrettable attrition of about 1% over the 12 months ending September 30. That measure covers the stated 12-month period; the report does not provide a definition of regrettable attrition or a breakdown by role.

The Times of India also said some teams at Anthropic had gone further in their use of AI to produce code, without specifying the teams or giving a separate figure for their output.

The company’s reported code contribution and the increase in engineer-level contributions are distinct measures: one describes Claude’s share of contributions in September, while the other compares daily contributions per engineer across periods. The figures and explanations are attributed to Anthropic in the account.

They present the company’s description of how it uses Claude in coding, model development and work on computing efficiency, alongside its stated aim of expanding without growing staff at the same pace.

Why it matters

  • Anthropic says Claude generated more than 90% of contributions to its codebase in September, making the company itself a case study in AI use for software development.
  • Anthropic reported that engineer-level daily code contributions in the July-to-September quarter were more than 45 times the 2024 average, while its workforce grew from about 940 to roughly 5,300.
  • The company says it is pursuing business growth without increasing staff at the same pace; its reported regrettable attrition was about 1% over the 12 months ending September 30.

What to watch next

  • Whether Anthropic reports further figures on Claude’s share of code contributions after September.
  • Whether Anthropic provides more detail about the teams it said had gone further in using AI to produce code.
  • Whether Anthropic reports how its workforce and regrettable attrition change as it pursues growth without adding staff at the same pace.

How it's being covered

Coverage analysisShow details
5 source articles · 5 unique outlets
Outlet political mix: Center Only· 4 Center · 1 unrated
Where the outlets are based: Unknown 60% · US/Canada 20% · Other 20%
Outlet political mix
Center Only
Left
Center
Right
0 outlets
4 outlets
0 outlets
The marker shows coverage balance: each unique outlet counts once. 1 unrated outlet is not included.
Geography
Where the outlets covering this story are based.
5 unique outletsDominant coverage: US/Canada
Coverage imbalance
  • Coverage is currently concentrated among Center-rated outlets (4 of 4 rated sources); comparatively few Left or Right-rated outlets have published on the event.
How we measure coverage

How outlets' accounts compare

Updat3 does not yet have enough full reports to compare how outlets framed this story.

Timeline

Developments from this story's sources, dated background we compiled and fact-checked, and our earlier coverage. Each row says which.
  1. Anthropic was founded in 2021 by former OpenAI employees, including Dario Amodei, who became its chief executive.
    Background · from our historical context
  2. GitHub and OpenAI announced Copilot, a tool that suggests code.
    Background · compiled and fact-checked by Updat3
  3. OpenAI’s Codex, the model behind Copilot’s early technical preview, had been introduced in August 2021.
    Background · from our historical context
  4. In December 2022, Anthropic researchers published “Constitutional AI,” a method for training assistants using written principles and AI feedback.
    Background · from our historical context
  5. Anthropic introduced Claude to the public in March 2023.
    Background · from our historical context
  6. Anthropic made Claude available to the public through Claude.ai.
    Background · compiled and fact-checked by Updat3
  7. The release followed the Claude 3 family, announced in March 2024, and gave Anthropic a model line explicitly positioned for software tasks.
    Background · from our historical context
  8. In June 2024, Anthropic released Claude 3.5 Sonnet, describing improvements in coding and introducing Artifacts, a feature for working with generated code and other content.
    Background · from our historical context
  9. Anthropic previewed computer use in Claude 3.5 Sonnet, enabling it to interact with computers through screenshots, mouse actions and keyboard input.
    Background · compiled and fact-checked by Updat3

How we got here

In June 2024, Anthropic released Claude 3.5 Sonnet, describing improvements in coding and introducing Artifacts, a feature for working with generated code and other content. The release followed the Claude 3 family, announced in March 2024, and gave Anthropic a model line explicitly positioned for software tasks.

Anthropic was founded in 2021 by former OpenAI employees, including Dario Amodei, who became its chief executive. In December 2022, Anthropic researchers published “Constitutional AI,” a method for training assistants using written principles and AI feedback. Anthropic introduced Claude to the public in March 2023.

Original sources

1 of 5 articles · US/Canada · Show all

Every story links to the reporting it was built from. How we choose sources

Spotted a factual error, a broken link or a misleading summary? Tell us.