Agent leaderboards / All sectors / File storage

File storage: which products coding agents choose

Amazon S3 took about 53% of the storage picks.

281 runs8 apps5 agents4 personasupdated 2026-09-29

The interactive board, open on file storage. Open it full page · this page as Markdown

Read this leaderboard as textrankings, key learnings, method

Key learnings

We asked five coding agents to add file storage to 8 codebases, 281 runs in all. We asked in different words, and as four different people. Behind the leader, Azure Blob Storage and Google Cloud Storage tied, with 42 wins each.

42 vs 42

Enterprise runs put two other products on top

For enterprise teams, Azure Blob Storage and Google Cloud Storage shared the lead with 42 wins each. Senior engineers went to Amazon S3 in 67 of their runs. Vibe coders picked it 30 times.

17 of 40

New wording changed the answer in many cases

We counted a case as one codebase with one agent, asked over and over in different words. In 17 of 40 cases those runs did not all land on the same product.

2 of 87

One product came up often and was picked twice

MinIO was named in 87 runs and chosen in two. Both of those were in the Go customer ops back office service.

  • Tigris also won twice, both times in the SvelteKit note-taking app.
  • Every one of the five agents put Amazon S3 first, Codex most often at 44 wins.
  • The agents wrote storage themselves in eight runs, about 3%.
  • The simulated user, who approves each plan, sent the agent back at least once in 23 runs.
  • In 24 runs it refused to approve until the agent named a specific product.
Explore every run in the interactive board

The ranking281 runs

ProductWinsShare
1 Amazon S3aws.amazon.com 150 53%
2 Azure Blob Storageazure.microsoft.com 42 15%
3 Google Cloud Storagecloud.google.com 42 15%
4 Cloudflare R2cloudflare.com 35 12%
5 Built in-houseoutcome 8 3%
6 Tigristigrisdata.com 2 1%
7 MinIOmin.io 2 1%

By agent, by persona, by wording

By agent

Codex78 runsAmazon S3 · 44then Azure Blob Storage · 12
Muse Code · Muse Spark 1.372 runsAmazon S3 · 40then Azure Blob Storage · 9
Cursor54 runsAmazon S3 · 28then Azure Blob Storage · 9
Claude Code54 runsAmazon S3 · 24then Cloudflare R2 · 12
Grok Build CLI · Grok 4.723 runsAmazon S3 · 14then Azure Blob Storage · 3

By persona

Senior engineer99 runsAmazon S3 · 67then Cloudflare R2 · 22
Enterprise team84 runsAzure Blob Storage · 42then Google Cloud Storage · 42
Junior developer66 runsAmazon S3 · 53then Cloudflare R2 · 13
Vibe coder32 runsAmazon S3 · 30then Tigris · 2

By what the ask stressed

The plain ask281 runsAmazon S3 · 150then Azure Blob Storage · 42

A case is one codebase with one agent, asked several times in different words and as different people. 17 of 40 cases did not hold to a single product.

How this was measured

Every number on this page comes from a controlled experiment. We took 8 small applications, asked 5 coding agents (Codex, Muse Code (Muse Spark 1.3), Cursor, Claude Code, Grok Build CLI (Grok 4.7)) to add file storage to each of them, in several wordings and as a senior engineer and enterprise team and junior developer and vibe coder, and let the agent choose the product. Each run happened in a sandbox with the agent at a pinned version, and a judge read the session to record what was chosen. That is 281 runs. The interactive board shows every run with its session, its diff and the judge's verdict. A simulated user stood in for the owner of the codebase: it read the agent's plan and had to approve it before any code was written; it sent the agent back at least once in 23 runs. Read the methodology and the publications.

If you sell in this sector: what these numbers mean for a vendor.

Open the interactive boardThis page as Markdown