Skip to main content

Benchmark profile

HAE-RAE Math 8K (HRM8K)

Korean mathematical reasoning (high-school to Olympiad level).

Data verified

How BenchLM shows HRM8K right now

BenchLM is tracking HRM8K in the local dataset, but exact-source verification records for these rows are still being attached. To avoid a blank benchmark page, BenchLM shows the current tracked rows below as a display-only reference table.

These tracked rows are useful for inspection and spot-checking, but until exact-source attachments are completed they should not be treated as fully verified public benchmark rows.

1 tracked modelsLocal tracked rowsAwaiting exact-source attachmentsDisplay only

Tracked score on HRM8K — July 23, 2026

BenchLM mirrors the published tracked score view for HRM8K. K-Exaone leads the public snapshot at 90.9%. BenchLM does not use these results to rank models overall.

1

K-Exaone

LG AI Research

k-exaone

90.9%
Overall
1 modelKorean BenchmarksKorean-language benchmarkRefreshingDisplay onlyUpdated July 23, 2026

Tracked score table (1 model)

Score
1
K-ExaoneLG AI Research
90.9%

About HRM8K

Tasks

8,011 instances

Format

Math word problems

Difficulty

Olympiad level

BenchLM freshness & provenance

Version

HRM8K

Refresh cadence

Static

Staleness state

Refreshing

Question availability

Public benchmark set

RefreshingDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does HRM8K measure?

Korean mathematical reasoning (high-school to Olympiad level).

Which model leads the published HRM8K snapshot?

K-Exaone currently leads the published HRM8K snapshot with 90.9% tracked score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on HRM8K?

1 AI models are included in BenchLM's mirrored HRM8K snapshot, based on the public leaderboard captured on July 23, 2026.

Last updated: July 23, 2026 · mirrored from the public benchmark leaderboard

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.