Checked fact 42758 Nov 2026Research

The paper's task suites total 170 tasks: HCAST (97 tasks from 46 families, about 1 minute to 30 hours), RE-Bench (7 machine-learning research engineering tasks of about 8 hours each) and SWAA (66 single-step tasks of 1 to 30 seconds) (section 2.1).

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

1. A subset of HCAST: 97 diverse software tasks ranging from 1 minute to 30 hours. ... 2. RE-Bench: 7 difficult ML research engineering tasks, all eight hours long. 3. Software atomic actions (SWAA): 66 single-step tasks representing short segments of work by software developers, ranging from 1 second to 30 seconds. ... We use 97 tasks from 46 task families in HCAST

What the source said when we opened it, on 8 Nov 2026.

The source

Measuring AI Ability to Complete Long Tasks (arXiv 2503.14499v3)
METR

Open the source archived copy

Checked

Checked by the notis newsroom on 8 Nov 2026, against the source above.

In the story

How METR's time horizon rates AI agents by human task length 8 Nov 2026

Cite this fact

Short permanent address

In a sentence

“The paper's task suites total 170 tasks: HCAST (97 tasks from 46 families, about 1 minute to 30 hours), RE-Bench (7 machine-learning research engineering tasks of about 8 hours each) and SWAA (66 single-step tasks of 1 to 30 seconds) (section 2.1).” ai notis, checked 8 Nov 2026. Source: Measuring AI Ability to Complete Long Tasks (arXiv 2503.14499v3) (METR). https://ainotis.com/c/4275

As a reference (APA style)

ai notis. (2026, November 8). The paper's task suites total 170 tasks: HCAST (97 tasks from 46 families, about 1 minute to 30 hours), RE-Bench (7 machine-learning research engineering tasks of about 8 hours each) and SWAA (66 single-step tasks of 1 to 30 seconds) (section 2.1) [Checked claim 4275]. ai notis. https://ainotis.com/c/4275

Anyone may quote this address. It does not change; if we correct the story, this page says so.