Evaluation & Benchmarks · 2025
Humanity's Last Exam
A 2,500-question multimodal exam written by subject-matter experts at the frontier of human knowledge, built as a successor benchmark after models saturated MMLU and to resist quick internet retrieval.
Editorial record
Plain-language summary
A 2,500-question multimodal exam written by subject-matter experts at the frontier of human knowledge, built as a successor benchmark after models saturated MMLU and to resist quick internet retrieval.
Knowledge graph
Relationships
Descendants
Conceptual ancestorEvidence: Strongly supported
MMLU: Massive Multitask Language Understanding
Builds on the evaluation lineage in the archive
freshness sweep 2026
Source record
Provenance
- Record ID
- P-619
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/2501.14249
- arXiv:2501.14249
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.