- GDPVal-AA-V2 (metric not stated, not an Elo) 28.5
- Terminal-Bench 2.1 51
- Tau3 Banking 17
- HLE (no tools) 25.3
- AA-Omniscience accuracy 16.90 (non-hallucination 61.27), not the net index 16.9
- SciCode 41
- GPQA Diamond 86.4
- AA-LCR 64.1
- PinchBench 87.8
- MCP-Atlas (500 public tasks) 52.2
- BrowseComp 66.8
- WebArena (vision) 50.9
Nemotron 3.5 Bolt 74B-A9B
Nemotron 3.5 Bolt 74B-A9B places in the top three on PinchBench 2.0.
12
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
NVIDIA Oct 2026 Nemotron 3.5 Bolt 74B-A9B
Open weights Mid-size
144.2
#138 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- IFBench 66.2% #52 /80
- SciCode 41% #27 /49
- AA-LCR (Artificial Analysis Long Context Reasoning) 64.1% #18 /38
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AA-LCR (Artificial Analysis Long Context Reasoning) | 64.1% | #18 of 38 | Lab model card | |
| Humanity's Last Exam | 25.3% | #92 of 154 | Lab model card | |
| IFBench | 66.2% | #52 of 80 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 86.4% | #64 of 302 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SciCode | 41% | #27 of 49 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| PinchBench 2.0 | 90.3% | #3 of 57 | Official leaderboard | |
| PinchBench | 87.8% | #6 of 14 | Lab model card | |
| BrowseComp | 66.8% | #52 of 81 | Lab model card | |
| τ³-bench Banking | 17% | #17 of 22 | Lab model card | |
| WebArena | 50.9% | #7 of 8 | Lab model card | |
| Terminal-Bench 2.1 | 51% | #51 of 56 | Lab model card | |
| MCP Atlas | 52.2% | #44 of 46 | Lab model card |
Its lab's numbers put it at 144.2, number 138 of 404
The Self-Reported Index fits one score to the 10 capability results NVIDIA reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 142.0 to 146.3. How we compute it.
- PinchBench 87.8
- GPQA Diamond 86.4
- BrowseComp 66.8
- IFBench 66.2
- AA-LCR (Artificial Analysis Long Context Reasoning) 64.1
- MCP Atlas 52.2
- Terminal-Bench 2.1 51
- SciCode 41
- Humanity's Last Exam 25.3
- τ³-bench Banking 17
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.
Its system card
Every result in the longer documents the lab published with it. 1 of them isn't in the launch post.
Capabilities
- GDPval-AA v2 28.5
- Terminal-Bench 2.1 51
- τ³-bench Banking 17
- Humanity's Last Exam 25.3
- SciCode 41
- GPQA Diamond 86.4
- AA-LCR (Artificial Analysis Long Context Reasoning) 64.1
- PinchBench 87.8
- MCP Atlas 52.2
- BrowseComp 66.8
- WebArena 50.9
- IFBench 66.2