Mi:dm K 2.5 Pro — benchmark results

Provider: KT. Released 2025-12-28. Access: API.

Unified ELO 1594 ± 35, rank #507 of 1841 rated models, from 53 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA TAU-2 Bench86.55Accuracy (%)79.6
AA LiveCodeBench65.61Pass@1 (%)73.4
AA MMLU-Pro80.94Accuracy (%)72.7
AA AIME 202576.67Accuracy (%)71.7
AA IFBench49.32Accuracy (%)58.7
AA GPQA Diamond70.1Accuracy (%)53.3
AA Humanity's Last Exam7.69Accuracy (%)53.3
Artificial Analysis Intelligence Index16.42Intelligence Index51
AA Omniscience - Science, Engineering & Mathematics26.2Accuracy (%)50.8
AA Omniscience - Software Engineering (SWE) - Rust50Accuracy (%)49.6
AA Global-MMLU-Lite - Korean83.5Accuracy (%)48.7
AA Global-MMLU-Lite - Italian87.25Accuracy (%)48.3

Interactive version: theaggregate.ai/model?slug=mi-dm-k-2-5-pro · How It Works · Data refreshed daily, snapshot 2026-07-25.