Skip to content

Nemotron 3 Ultra

NVIDIAOpen weightsReasoningProvisionalReleased Jun 4, 2026Input: text
Compare with Ling 3.0 Flash VLSource scores ↗OpenRouter ↗
Overall score
46.8
Rank 117 of 222 · 32.4–61.1
Context window
1M
1,000,000 tokens
Price per 1M tokens
$0.60 in · $2.40 out
Source: OpenRouter
Blended price
$1.05
3:1 input:output tokens
Output speed
—
Not measured yet

Category scores

Scores per capability area with the model's rank among ranked models in that category.

  1. 1Instruction following#2092.1
  2. 2Multilingual#789.2
  3. 3Reasoning69.5
  4. 4Coding#11758.9
  5. 5Knowledge#8457.4
  6. 6Agentic & tool use#10120.4
Higher is better. #N is the model's rank in that category, where it is eligible to be ranked.

Benchmark results (18)

As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.

Agentic & tool use

Terminal-Bench 2.156.4#55 of 68
BrowseComp44.4#48 of 50
MazeBench0#20 of 42
HLE w/ tools37.4#25 of 27
τ³-bench results70.9#4 of 23
PinchBench90.0#2 of 7

Knowledge

GPQA87.0#40 of 117
HLE26.7#50 of 69
MMLU-Pro86.8#9 of 67
HLE w/o tools26.7#35 of 43

Coding

SWE-bench Verified71.9#57 of 98
LiveCodeBench v689.0#10 of 46
SciCode44.6#11 of 23

Multilingual

SWE Multilingual67.7#37 of 46
MMLU-ProX83.0#7 of 16

Reasoning

LongBench v261.9#6 of 17

Multimodal

Design Arena Website1145#74 of 91

Instruction following

IFBench81.7#4 of 42

Pricing by source

USD per 1M tokens.

SourceInputOutputCached in
List price (OpenRouter)$0.60$2.40—
OpenRouter nvidia/nemotron-3-ultra-550b-a55b$0.60$2.40$0.12

Open-weight model. Self-hosted or third-party hosted costs vary by provider and infrastructure.

Open weights

From the Hugging Face repository.

Parameters
561B
Licence
other
Downloads (30 days)
1M

Speed

Measured by us through OpenRouter (median of trials).

Not measured yet.

Related models

Similar score