Ox Alpha (Max): benchmark results

Provider: Other. Released 2026-08-20.

Unified ELO 1720 ± 27, rank #328 of 2075 rated models, from 28 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
DuelLab Overall57DuelLab Score86.5
LiveBench Table Reformat100Score81.5
K-Bench98.36Overall (%, Default v1 prompt)81.1
LiveBench Python60Score72.3
K-Bench (Therapeutic v0)97.84Overall (%, Therapeutic v0 prompt)71
LiveBench AMPS Hard98Score58.5
K-Bench - Risk92.26Risk Score (%, Default v1 prompt)57.4
LiveBench TypeScript43.33Score55.4
K-Bench - Risk (Therapeutic v0)91.2Risk Score (%, Therapeutic v0 prompt)54
LiveBench Code Completion78.26Score48.5
LiveBench Consecutive Events80.35Score47.7
LiveBench Story Generation70.87Score47.7

Interactive version: theaggregate.ai/model?slug=ox-alpha-max · How It Works · Data refreshed daily, snapshot 2026-10-04.