Nemotron 3 Ultra 550B A55B
nvidia/nemotron-3-ultra-550b-a55b
Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy
Capabilities
ReasoningYes
Tool callingYes
Structured outputNo
TemperatureSupported
AttachmentNo
Open weightsYes
Limits
Context1,000,000 tokens
Max inputN/A
Max output128,000 tokens
Details
Familynemotron
Knowledge cutoffN/A
Release date2026-06-04
Last updated2026-06-04
Modalities
Inputtext
Outputtext
Benchmarks
SWE-Bench Verified
70.7
resolved
SWE-Bench Multilingual
67.7
resolve rate
Terminal-Bench
56.4
success rate
GPQA
87
accuracy
Humanity's Last Exam
26.7
accuracy
Humanity's Last Exam
37.4
accuracy
LiveCodeBench
89
pass@1
MMLU-Pro
86.8
accuracy
BrowseComp
44.4
accuracy
IFBench
81.7
accuracy
GDPval
46.7
wins or ties
Community Rating
—
No ratings yet
Sign in to rate
Pricing
Cost per 1M tokens
InputN/A
OutputN/A