VECTOR WIREAI INTELLIGENCE
⌘K
PKT
Refresh Models —Deals —Regulatory —Sources —

Questions this page answers

Which is better, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 or NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16?
Across 11 shared benchmarks, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 scores higher on 2 and NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 on 9. The widest gap is MMLU-Pro, where NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 scores 79.1 against 67.9.

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 vs NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16

Across 11 shared benchmarks, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 scores higher on 2 and NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 on 9. The widest gap is MMLU-Pro, where NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 scores 79.1 against 67.9.

NVIDIAvsNVIDIA11 shared benchmarks2–9 head-to-head
BenchmarkNVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16
AGIEval-En7078.7
arc_challenge92.797.3
GSM8K91.388.1
hellaswag85.590.5
humaneval77.483.8
mmlu78.689.1
MMLU-Pro67.979.1
OpenBookQA47.648.6
RULER 1M69.676.8
RULER 256K76.986.2
winogrande8079.3

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.