NVIDIA·released 2026-06-041 source
Nemotron 3 Ultra benchmark scores: 4 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 4 benchmarks·best result 43.45 on WeirdML (unverified)·0 of 4 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| WeirdML | 43.45accuracy (%) | unverifiedT1 | 2026-06-04https://htihle.github.io/weirdml.html |
| APEX-Agents | 11.5accuracy (%) | unverifiedT2 | 2026-06-04 |
| CritPt | 3.14accuracy (%) | unverifiedT2 | 2026-06-04 |
| ProofBench | 2accuracy (%) | unverifiedT1 | 2026-06-04https://www.vals.ai/benchmarks/proof_bench |
Claims drawn from cited facts, not live model generation.
Developed by NVIDIA, this model was released in 2026.
2 cited facts
This model is tracked on four benchmarks and currently holds no top score on any of them. Because its record is unverified, neither confirmed nor disputed by outside evaluation, these figures should be treated with caution.
3 cited facts
Relative to disclosed-harness leaderboard scores, this model trails the average SOTA score by 52.04 points, which reads as a wide gap and places it well behind the leaders. It also holds the top score on none of the tracked benchmarks, confirming no category leadership yet.
2 cited facts
1 facts cross-checked across data sources: 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks
Nemotron 3 Ultra is an AI model developed by NVIDIA, released 2026-06-04. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Nemotron 3 Ultra has recorded scores on 4 benchmarks, each shown with its evidence status.
Nemotron 3 Ultra has recorded scores on 4 benchmarks — WeirdML, CritPt, APEX-Agents, ProofBench. The full table above shows each score with its evidence status.
0 of 4 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
Nemotron 3 Ultra has 4 tracked claims: 4 unverified.