Benchmark Almanac

Nemotron-Cascade 2 30B-A3B

Nemotron-Cascade 2 30B-A3B ranks first on 2 of the 26 benchmarks with five or more models scored.

NVIDIA Released March 19th, 2026 Family Nemotron-Cascade The lab's model card Compare with others
28
Benchmarks with a score
2
Ranked first, out of five or more models
5
In the top three
NVIDIA Mar 2026

Nemotron-Cascade 2 30B-A3B

Open weights Small

138.5 #191 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • IFBench 82.9% #1 /80
  • SciCode 36.4% #38 /49
  • AA-LCR (Artificial Analysis Long Context Reasoning) 39.1% #34 /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

BenchmarkScoreRankSource
IFBench 82.9% #1 of 80 Lab model card
RULER 99% #1 of 42 Lab model card
Arena-Hard v2.0 83.5% #3 of 24 Lab model card
MMLU-Pro 79.8% #82 of 167 Lab model card
WMT24++ 84.1% #9 of 15 Lab model card
MMLU-ProX 72.5% #16 of 23 Lab model card
Humanity's Last Exam 17.7% #116 of 154 Lab model card
MMLU-Redux 86.3% #34 of 45 Lab model card
MultiChallenge 45.3% #33 of 42 Lab model card
AA-LCR (Artificial Analysis Long Context Reasoning) 39.1% #34 of 38 Lab model card
LongBench v2 40.3% #32 of 35 Lab model card
CL-bench 12.2% #6 of 6 Lab model card

Math and science

BenchmarkScoreRankSource
HMMT 2025 94.6% #14 of 54 Lab model card
AIME 2025 92.4% #38 of 143 Lab model card
IMO-ProofBench 72.9% #2 of 5 Lab model card
GPQA Diamond 76.1% #143 of 302 Lab model card
IMO 2025 83.3% #3 of 6 Lab model card
IMO-AnswerBench 79.3% #25 of 37 Lab model card
AIME 2026 90.9% #27 of 38 Lab model card

Coding

BenchmarkScoreRankSource
LiveCodeBench v6 87.2% #14 of 90 Lab model card
Codeforces rating 2,320 #13 of 47 Lab model card
ICPC World Finals 2025 10.0 #2 of 3 Lab model card
SciCode 36.4% #38 of 49 Lab model card
SWE-bench Verified 50.2% #128 of 155 Lab model card
IOI 2025 439.3 #3 of 3 Lab model card

Agents and tools

BenchmarkScoreRankSource
BFCL V4 52.9% #25 of 30 Lab model card
tau2-bench 58.9% #66 of 79 Lab model card
Terminal-Bench 2.0 21.1% #62 of 63 Lab model card

Its lab's numbers put it at 138.5, number 191 of 404

The Self-Reported Index fits one score to the 22 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 135.7 to 141.2. How we compute it.

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.