Can AI finish a long coding task alone?

The answer

Stage 3: independently tested.

The rule that placed itSomeone other than the maker has tested it: an independent benchmark, a peer-reviewed paper, or a hands-on test by a trusted outlet.

The evidenceMETR's May 2026 risk report says the most capable agents it evaluated had a measured time horizon of over two full working days, but it is increasingly unsure of the figure because its tests are saturating. METR, Frontier Risk Report (February to March 2026), May 19, 2026.

A stage says how far the evidence has got. It is not a rating of the AI and not advice.

1
ClaimedEvidence found
2
LaunchedNo evidence yet
3
Independently testedPlaced here
4
Tested with real peopleNo evidence yet
5
Regulator-clearedNo approval process exists
6
Widely used with a verified recordNo evidence yet

The evidence

Everything we found.

In ladder order, newest first within each kind. Makers' own claims are labeled as theirs.

Claimed

What the makers claim

Anthropic (company-reported)
“One tester completed a 680,000-line code migration in less than a day—work that would have taken an engineering team weeks.”
Anthropic, Introducing Claude Opus 5.5 · Sep 22, 2026 · company-reported

Tested

What independent tests found

METR
METR's May 2026 risk report says the most capable agents it evaluated had a measured time horizon of over two full working days, but it is increasingly unsure of the figure because its tests are saturating.
“Their measured time horizon was over two full-time-equivalent days”
METR, Frontier Risk Report (February to March 2026) · May 19, 2026
METR
In METR's public data, the best model finishes, half the time, software tasks that take a skilled person about 12 hours. METR says its tests can't measure reliably above 16.
METR, Time Horizon 1.1 · Feb 5, 2026

Not found in our sources yet: what has launched; tested with real people; how many people use it. No regulator approves AI for this task, so the approval step doesn't apply.

Beside the stage