- AA-Omniscience Accuracy14%#188
- Humanity's Last Exam2.9%#261
- AA-LCR18.7%#199
- CritPt0%#143
- MMMU-Pro49%#104
- GPQA Diamond60.3%#143
- IFBench40.5%#86
- τ²-Bench Telecom62%#57
- Terminal-Bench Hard6.8%#71
Nova 2.0 Lite (Non-reasoning)Median
Data from Artificial Analysis