16 parallel Claude agents wrote a C compiler that builds the Linux kernel
An Anthropic researcher ran a team of Claude agents in a loop on a shared repository to write a Rust-based C compiler from scratch. The write-up and the code are both public.
The problem
It tests whether agents working mostly without supervision can finish a large, tightly specified engineering project when good tests and an oracle compiler keep them honest.
How they did it
- Several Claude instances each run in an infinite loop, picking the next task by themselves.
- Git and lock files keep agents from working on the same thing.
- Strong test suites, CI and GCC as a reference oracle tell the agents whether their output is correct.
Results
- Over nearly 2,000 Claude Code sessions and $20,000 in API costs, the agents produced a 100,000-line compiler.
- It can build a bootable Linux 6.9 on x86, ARM, and RISC-V.
- It reaches a 99% pass rate on most compiler test suites and also compiles QEMU, FFmpeg, SQLite, Postgres and Redis.
- It is not a drop-in replacement for GCC. It relies on GCC for 16-bit x86 code and produces less efficient code.
As reported by the source (Anthropic Engineering); AgentGid did not measure these figures.
This suits a well-funded team with strong engineering skills that already has a mature test suite and a reference implementation to check against. The $20,000 in API costs is far beyond the $20/mo Claude Code plan price, so budget for pay-as-you-go usage. Without an oracle like GCC, the agents have nothing to keep them honest. For a cheaper way to try the loop on a smaller project, Pi (Free (OSS)) charges only model provider costs.
The agent used here
Similar use cases
Turn a Figma frame into code by connecting the Figma MCP server
Figma's hosted MCP server gives a coding agent the variables, components and layout data of a frame you link to, so the code it writes matches the design.
The agent gets design context such as variables, components and layout for the selected frame.
How Rakuten ran a seven-hour autonomous vLLM change with Claude Code
Rakuten engineers use Claude Code across the development lifecycle. In one test, it implemented a method in the open-source vLLM library in a single unattended run.
As reported by Anthropic: 'Claude Code finished the entire job in seven hours of autonomous work in a single run.'
How LG CNS migrated a 20-year-old system with a Claude Code build pipeline
LG CNS built an orchestration layer around Claude Code, which it calls a Build Factory, and used it to modernize a construction company's 20-year-old project management system. As…
As reported by Anthropic: 2,888 of 2,913 APIs converted, 'a 99.1% completion rate'.
Guides
Best AI Agents in 2026: Live Rankings by Category
The leading AI agents for coding, design, research, data, marketing, sales, support, HR, finance, legal, DevOps and security, ranked by public usage data that refreshes daily, with prices and free options.
What Is an AI Agent? A Practical Explanation With Real Examples
What makes software an AI agent, how agents differ from chatbots and plain language models, the main types in use today and how to judge one.
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.
Best AI Coding Agents in 2026: Ranked by Usage, Benchmarks and Price
The AI coding agents developers use most, with live download and install numbers, benchmark results, entry prices and a way to choose between terminal, IDE and cloud agents.