Agentic AI Atlasby a5c.ai
/
GitHubDocsDiscord
iiRecord
Agentic AI Atlas · RE-Bench AI R&D task suite
test-set:re-bench-ai-rd-tasksa5c.ai
II.
TestSet overview

test-set:re-bench-ai-rd-tasks

Reference · live

RE-Bench AI R&D task suite overview

Canonical RE-Bench artifact for frontier AI R&D agent evaluation.

TestSetOutgoing · 1Incoming · 0

Attributes

displayName
RE-Bench AI R&D task suite
benchmarkId
benchmark:re-bench
caseCount
7
composition
RE-Bench's staged task suite measures autonomous AI research and development work against human-expert baselines using realistic machine-learning research-engineering tasks.
homepageUrl
https://github.com/METR/RE-Bench
description
Canonical RE-Bench artifact for frontier AI R&D agent evaluation.

Outgoing edges

belongs_to_benchmark1

Incoming edges

None.