benchgap
NVIDIA · model

Nemotron Ultra 253B benchmark scores

As of 2026-10-07, Nemotron Ultra 253B (NVIDIA) has measured scores on 7 benchmarks and estimated scores on 33 more.

BenchmarkScoreSource
AA-GPQA Diamond72.8%measured
AA-HLE7.4%measured
Artificial Analysis Intelligence Index7.5%measured
CritPt0.0%measured
GDPval-AA0.0%estimated ± 12.5 pp, low confidence
GPQA Diamond71.9%estimated ± 2.2 pp, high confidence
HLE19.0%estimated ± 9.5 pp, medium confidence
IFBench55.7%estimated ± 6.9 pp, low confidence
AA-Omniscience Accuracy20.1%measured
IFEval87.7%estimated ± 1.3 pp, high confidence
τ²-bench results11.4%measured
APEX-Agents-AA1.6%estimated ± 11.4 pp, low confidence
BioMysteryBench (human-difficult)27.2%estimated ± 5.9 pp, low confidence
BFCL v444.9%estimated ± 9.8 pp, low confidence
AA Agentic Index0.0%estimated ± 12.3 pp, low confidence
MMLU-Pro85.0%estimated ± 3.5 pp, high confidence
GPQA48.1%estimated ± 3.2 pp, medium confidence
AA-IFBench38.2%measured
AA EnterpriseOps-Gym1.2%estimated ± 3.6 pp, low confidence
ARC-AGI-10.6%estimated ± 5.4 pp, low confidence
ARC-AGI-20.3%estimated ± 14.2 pp, low confidence
Vals GPQA Diamond62.0%estimated ± 7.4 pp, medium confidence
Vals MMLU-Pro83.1%estimated ± 2.2 pp, high confidence
AA ITBench100.0%estimated ± 3.6 pp, low confidence
HLE w/o tools5.9%estimated ± 5.7 pp, low confidence
C-Eval79.0%estimated ± 1.0 pp, low confidence
MCP-Tasks100.0%estimated ± 8.2 pp, low confidence
MMLU-Redux93.3%estimated ± 1.1 pp, medium confidence
SuperGPQA64.3%estimated ± 7.5 pp, medium confidence
HealthBench Hard0.0%estimated ± 7.7 pp, low confidence
MMLU-Pro (Arcee)77.1%estimated ± 4.0 pp, low confidence
ARC-AGI-31.7%estimated ± 3.4 pp, medium confidence
BioMysteryBench (human-solvable)81.7%estimated ± 0.6 pp, low confidence
HealthBench (raw)11.4%estimated ± 6.4 pp, low confidence
HealthBench (length-adjusted)53.9%estimated ± 3.7 pp, low confidence
HealthBench Professional57.7%estimated ± 4.8 pp, low confidence
HealthBench Professional (raw)25.1%estimated ± 5.0 pp, low confidence
HLE-Verified15.3%estimated ± 2.0 pp, low confidence
LABBench253.9%estimated ± 1.8 pp, low confidence
MMLU86.3%estimated ± 1.1 pp, medium confidence