A sunlit acoustic studio with an amplifier, measurement microphones, a recording console, and a small potted aloe by the right wall.

Understand how AI agents choose, build, and perform.

Benchmarks and research on the choices agents make, the code they write, and the tasks they complete.

Record what coding agents do. See what to improve.

Used by leading enterprise DevTools and infrastructure companies.

From the researchers behind the original What Claude Code Chooses study.

Latest research / October 1, 2026

What Dots Actually Choose: OpenAI’s Agent Already Has a Stack

All 8 builds used Cloudflare Workers and D1 through ChatGPT Sites, with ChatGPT sign-in. Follow the builds, see the apps, and compare their approaches to scheduled scraping.

Read the research

Browse all research Explore the Coding Agents Index

Two products

See what agents choose.

Four distinct amplifier cabinets arranged for comparison in a bright acoustic gallery, each with a measurement microphone. — The lineup: compare how coding agents choose between products.

01 / Choices, findings, playbooks

Benchmarking

Measure when coding agents choose your product, what they choose instead, and why. Get findings and prioritized playbooks for your team.

Product selection and competitive benchmarks
Sources and evidence behind each finding
Playbooks for docs, positioning, product features, and developer experience

Explore Benchmarking

Test what helps them succeed.

A substantial analog recording console overlooking one amplifier and microphone through a test chamber window. — The control room: change the inputs and measure the difference.

02 / Changes, tests, results

Ground Control

Make a change to your docs, skills, or tooling. Rerun the same task to see what gets better or worse, then iterate.

Installation, authentication, and real workflows
Recorded runs with inspectable tool calls
Before-and-after tests of docs and tooling

Explore Ground Control

Tool choice and security.

Browse all research

Our original study / February 2026

We pioneered research on what coding agents choose.

Our original What Claude Code Chooses study put numbers behind agents’ developer-tool decisions. Today, we help developer tool teams measure and improve how coding agents find, understand, and use their products.

Read the original study

Safety / April 2026

What security defaults do agents choose?

Six development tasks leave the security defaults unspecified. The report checks the resulting code against 33 exploit tests.

Read study

Get research updates

New Amplifying research on coding agents, delivered by email.

See Amplifying in action.

Tell us what you’re working on. We’ll show you where Amplifying can help.

Set up a call