Apodex 1.1: benchmark results

Apodex's proprietary reasoning-first agentic model (API and web workbench, size undisclosed); the open Apache-2.0 Apodex 1.1 Mini is a separate 35B-A3B checkpoint. Provider: Other. Released 2026-08-24. Access: API.

Unified ELO 1879 ± 47, rank #75 of 2656 rated models, from 37 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
BenchmarkList ECI137.81Capability Index (ECI)87.5
Artificial Analysis Intelligence Index30.37Intelligence Index87.1
AA Long Context Reasoning79.33Accuracy (%)86.4
AA Humanity's Last Exam34.06Accuracy (%)85.6
AA MMMU-Pro79.19Accuracy (%)83.8
AA-Omniscience Index - Software Engineering (SWE) - Rust56Omniscience Index83.8
AA GPQA Diamond86.36Accuracy (%)80.8
AA-LCR74.67Accuracy (self-reported)79
AA CritPt4.57Accuracy (%)76.5
AA-Omniscience Index - Science, Engineering & Mathematics-3.2Omniscience Index76.4
AA-Omniscience Accuracy31.68Accuracy (%)75.8
AA-Omniscience Index - Software Engineering (SWE) - HTML18Omniscience Index74.3

Interactive version: theaggregate.ai/model?slug=apodex-1-1 · How It Works · Data refreshed daily, snapshot 2026-09-19.