- AA-Omniscience Accuracy16.9%#151
- Humanity's Last Exam3.9%#246
- AA-LCR30%#174
- CritPt0%#143
- GPQA Diamond63.6%#134
- IFBench52%#57
- τ²-Bench Telecom71.6%#49
- Terminal-Bench Hard16.7%#45
Nova 2.0 Pro Preview (Non-reasoning)Median
Data from Artificial Analysis