NVIDIA · model
Nemotron 3.5 Bolt 74B-A9B benchmark scores
As of 2026-10-29, Nemotron 3.5 Bolt 74B-A9B (NVIDIA) has measured scores on 12 benchmarks and estimated scores on 85 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 86.0% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 24.7% | measured |
| AA-SciCode | 39.7% | measured |
| CritPt | 4.3% | measured |
| GDPval-AA | 24.0% | measured |
| GPQA Diamond | 86.4% | measured |
| HLE | 30.4% | estimated ± 6.6 pp, medium confidence |
| AA-LCR | 69.7% | measured |
| LiveCodeBench v6 | 80.4% | estimated ± 4.9 pp, high confidence |
| SciCode | 39.9% | estimated ± 3.1 pp, high confidence |
| Terminal-Bench 2.1 | 51.0% | measured |
| BrowseComp | 66.8% | measured |
| HLE w/ tools | 45.4% | estimated ± 5.3 pp, low confidence |
| LongBench v2 | 60.2% | estimated ± 5.8 pp, low confidence |
| VITA-Bench | 37.9% | estimated ± 6.7 pp, low confidence |
| τ²-bench results | 91.4% | estimated ± 11.5 pp, low confidence |
| AA Coding Index | 36.7% | estimated ± 6.1 pp, medium confidence |
| APEX-Agents | 11.6% | estimated ± 0.6 pp, low confidence |
| APEX-Agents-AA | 14.4% | estimated ± 4.2 pp, high confidence |
| BioMysteryBench (human-difficult) | 35.5% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 78.6% | estimated ± 7.3 pp, low confidence |
| SWE-bench Verified | 65.5% | estimated ± 6.4 pp, medium confidence |
| AutomationBench | 20.7% | estimated ± 10.5 pp, low confidence |
| BFCL v4 | 73.2% | estimated ± 2.7 pp, high confidence |
| CyberGym | 50.9% | estimated ± 6.0 pp, low confidence |
| JobBench | 26.8% | estimated ± 6.4 pp, medium confidence |
| SWE-bench Pro | 43.3% | estimated ± 6.6 pp, medium confidence |
| τ³-bench results | 17.0% | measured |
| WideResearch | 70.9% | estimated ± 3.5 pp, high confidence |
| AA Agentic Index | 18.3% | estimated ± 2.9 pp, high confidence |
| MMLU-Pro | 84.9% | estimated ± 2.5 pp, medium confidence |
| Artificial Analysis Intelligence Index | 20.8% | estimated ± 5.8 pp, medium confidence |
| GPQA | 86.3% | estimated ± 1.2 pp, high confidence |
| AA-Omniscience Accuracy | 29.3% | estimated ± 7.8 pp, medium confidence |
| MMMLU | 88.2% | estimated ± 0.7 pp, medium confidence |
| AA MMLU-Pro | 85.9% | estimated ± 1.5 pp, high confidence |
| Gert Labs | 46.5% | estimated ± 7.4 pp, medium confidence |
| AA EnterpriseOps-Gym | 33.4% | estimated ± 3.3 pp, low confidence |
| AA Harvey LAB v1.0 | 87.2% | estimated ± 9.8 pp, low confidence |
| AA Tau3 Banking | 26.6% | estimated ± 3.9 pp, medium confidence |
| AA Terminal-Bench 2.1 | 60.2% | estimated ± 6.8 pp, low confidence |
| ARC-AGI-1 | 83.1% | estimated ± 2.7 pp, medium confidence |
| ARC-AGI-2 | 43.6% | estimated ± 10.1 pp, low confidence |
| CursorBench 3.2 | 37.2% | estimated ± 5.3 pp, low confidence |
| Terminal-Bench 3.0 | 3.7% | estimated ± 9.1 pp, low confidence |
| FrontierCode 1.1 Main | 0.0% | estimated ± 6.6 pp, low confidence |
| FrontierSWE v2 | 5.2% | estimated ± 14.6 pp, low confidence |
| MazeBench | 0.4% | estimated ± 5.4 pp, low confidence |
| OSWorld-Verified | 60.4% | estimated ± 5.8 pp, medium confidence |
| Vals GPQA Diamond | 85.6% | estimated ± 2.0 pp, high confidence |
| Vals MMLU-Pro | 84.6% | estimated ± 1.6 pp, high confidence |
| Vals SWE-bench | 66.0% | estimated ± 7.5 pp, low confidence |
| Terminal-Bench 2.1 (Vals AI) | 45.3% | estimated ± 7.4 pp, low confidence |
| AA AutomationBench | 8.9% | estimated ± 14.8 pp, low confidence |
| CWE-Bench-AA | 34.7% | estimated ± 5.8 pp, medium confidence |
| AA Cyber Index | 33.0% | estimated ± 13.3 pp, low confidence |
| DeepsecBench-AA | 0.8% | estimated ± 8.6 pp, low confidence |
| GDP.pdf | 6.5% | estimated ± 4.9 pp, low confidence |
| AA ITBench | 39.7% | estimated ± 4.1 pp, high confidence |
| AA Terminal-Bench 4.0 | 1.4% | estimated ± 11.6 pp, low confidence |
| CWE-bench v1 | 33.4% | estimated ± 6.4 pp, low confidence |
| DeepSWE | 29.8% | estimated ± 7.5 pp, low confidence |
| HLE w/o tools | 25.3% | measured |
| PostTrainBench v1.1 | 25.5% | estimated ± 7.4 pp, low confidence |
| Toolathlon-Verified | 49.9% | estimated ± 3.1 pp, medium confidence |
| ARC-AGI-3 | 0.0% | estimated ± 3.6 pp, medium confidence |
| ARC-AGI-3 Provider Adapter | 2.8% | estimated ± 7.5 pp, low confidence |
| Hermes Bench | 54.8% | estimated ± 2.9 pp, high confidence |
| Hermes Index | 22.1% | estimated ± 8.7 pp, medium confidence |
| ExploitGym | 0.8% | estimated ± 10.9 pp, low confidence |
| C-Eval | 95.0% | estimated ± 1.1 pp, low confidence |
| Claw-Eval | 63.4% | estimated ± 6.3 pp, low confidence |
| DeepPlanning | 35.8% | estimated ± 13.1 pp, low confidence |
| MCP Atlas | 52.2% | measured |
| MCP-Tasks | 75.4% | estimated ± 0.6 pp, low confidence |
| MMLU-Redux | 92.1% | estimated ± 2.8 pp, high confidence |
| NL2Repo | 31.6% | estimated ± 7.4 pp, medium confidence |
| QwenClawBench | 54.3% | estimated ± 2.6 pp, high confidence |
| SuperGPQA | 70.5% | estimated ± 13.9 pp, low confidence |
| Toolathlon | 37.2% | estimated ± 4.2 pp, high confidence |
| HealthBench Hard | 22.4% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (Text) | 44.7% | estimated ± 4.5 pp, low confidence |
| MMLU-Pro (Arcee) | 85.6% | estimated ± 3.4 pp, medium confidence |
| React Native Evals | 72.3% | estimated ± 4.7 pp, medium confidence |
| ResearchClawBench | 16.4% | estimated ± 1.8 pp, medium confidence |
| BioMysteryBench (human-solvable) | 81.8% | estimated ± 0.6 pp, low confidence |
| DRACO | 71.4% | estimated ± 3.1 pp, medium confidence |
| HealthBench (raw) | 33.2% | estimated ± 6.0 pp, low confidence |
| HealthBench (length-adjusted) | 47.8% | estimated ± 3.4 pp, low confidence |
| HealthBench Professional | 58.1% | estimated ± 4.5 pp, low confidence |
| HealthBench Professional (raw) | 44.8% | estimated ± 4.5 pp, medium confidence |
| STEMBench | 2.4% | estimated ± 0.2 pp, medium confidence |
| Agents' Last Exam | 15.8% | estimated ± 13.0 pp, low confidence |
| OpenHarmony Bench | 51.0% | estimated ± 2.4 pp, medium confidence |
| MRCRv2 | 72.9% | estimated ± 11.3 pp, low confidence |
| MMLongBench-Doc | 60.9% | estimated ± 2.6 pp, low confidence |
| PinchBench | 87.8% | measured |