263 projects tested, and "No Rust toolchain installed" stopped 9: week of August 31, 2026

What it ran on

Argusic publishes other people's server prices every day. These are its own: the machines this week's work was rented on, at the rates actually paid, each one destroyed when its job finished.

Attempts on CPU machines

482 at 0.1053 USD an hour

Attempts on GPU machines

2 at 0.6150 USD an hour

Data centers used

CA-MTL-1, EU-RO-1, EUR-IS-1, US-CA-2, US-KS-2, US-MD-1, US-MO-1, US-NC-1

What the week cost, broken down

A single total answers nothing for somebody working out whether they could do this themselves. The difference between the dearest and the cheapest project is how much fixing each needed, not a price anybody chose.

Model tokens

118.00 USD

Machines

36.86 USD

Per project measured

0.59 USD

Dearest project

deepseek-harness at 6.50 USD

Cheapest project

magic-mcp at 0.01 USD

What the pool looks like now

The state of the whole body of evidence at the end of this week, not only what the week itself added.

Projects measured in total

271

Of those, reached a working state

176

Most measured languages

Python (71), TypeScript (60), Rust (38), Go (36), JavaScript (26)

What kept breaking

Counted across projects, not across attempts, so one repository retried three times is one problem. Each label is the wording an agent actually used.

9 projects

No Rust toolchain installed

2 projects

2 test failures

2 projects

Go not installed in container

2 projects

No C compiler (cc/gcc) installed in container

2 projects

No Ruby installed in container

The result most worth reading

astron-rpa scored 100 of 100, after hitting 11 errors that the agent fixed with no human help.

Engine Python 3.13.15 dependencies installed, engine modules import correctly, 207 pytest tests passing across data processing and encryption components, 1 source bug fixed (string fill slice type). Frontend shared and CLI packages build. Can't run backend services or Electron desktop app without Windows/Docker.

The agent's own words, from its recorded report.

What this week does not tell you

Behavior over time

each attempt uses the smallest documented settings and a fixed time budget, so this measures whether a thing runs, not how it holds up under load

Anything not tested

only the 263 projects above were run this week; a project absent from this report was not measured, not judged

Real-service behavior for 84 of them

those ran against stand-in services, never the real ones

The report as written

Argusic installed and ran 263 projects on clean machines between 2026-08-31 and 2026-09-07, over 381 attempts, 252 more subjects than the week before.

171 were verified against the real thing, 84 only against stand-in services, and 126 could not be verified at all.

260 projects were tested here for the first time.

The same problems keep appearing across unrelated projects. The most common this week was "No Rust toolchain installed", which stopped 9 projects. Also recurring: "2 test failures" (2), "Go not installed in container" (2), "No C compiler (cc/gcc) installed in container" (2), "No Ruby installed in container" (2).

The result most worth reading was astron-rpa, which scored 100 of 100 after hitting 11 errors that the agent fixed with no human help. In the agent's own words: Engine Python 3.13.15 dependencies installed, engine modules import correctly, 207 pytest tests passing across data processing and encryption components, 1 source bug fixed (string fill slice type). Frontend shared and CLI packages build. Can't run backend services or Electron desktop app without Windows/Docker.

Running all of this cost 154.86 USD, of which 76% was model tokens rather than machines.

118.00 USD of that was model tokens and 36.86 USD was machines, which is 0.59 USD for each project measured.

The dearest project of the week was deepseek-harness at 6.50 USD, the cheapest magic-mcp at 0.01 USD. The difference is how much fixing each one needed, not a price we chose.

The work ran on machines rented by the hour and destroyed afterwards: 482 attempts on CPU machines at 0.1053 USD an hour, and 2 attempts on GPU machines at 0.6150 USD an hour. Those are the rates Argusic actually paid, not list prices. The pods came up in 8 data centres: CA-MTL-1, EU-RO-1, EUR-IS-1, US-CA-2, US-KS-2, US-MD-1, US-MO-1, US-NC-1.

Across everything measured so far, 271 projects have been installed and run here, and 176 of them reached a working state. Most of them are written in Python (71), TypeScript (60), Rust (38), Go (36), JavaScript (26).

Every number above comes from a recorded run, and each one has its log and its terminal recording on this site.