- AA-Omniscience Accuracy19%#122
- Humanity's Last Exam11.6%#134
- AA-LCR60.3%#128
- CritPt0.3%#117
- MMMU-Pro63.8%#80
- GPQA Diamond81.1%#68
- IFBench70.8%#18
- τ²-Bench Telecom72.8%#46
- Terminal-Bench Hard16.7%#45
Nova 2.0 Lite (High)Median
Data from Artificial Analysis