Nemotron 3 Ultra
NVIDIA · 1 configuration
Intelligence22.9
Input / 1M$0.60
Output / 1M$2.50
Context262.1K
Tariff: NVIDIA · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost1 measured configurations · 0 without both values
Selected: thinking thinking · Click a point to switch configuration
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
22.9 pointsGPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
86.7 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
28.4 %SciCodeCode generation for scientific research problems.
40.3 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
53.9 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
83.3 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
81.4 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
22.6 %Specifications from the catalog
Inputtext
Outputtext
Tool callingYes
Structured outputYes
Speed206.9 tok/s
First token1.05 s
Maximum output182.5K
Catalog listing date2026-06-04