Nemotron 3 Ultra Benchmark Details

Nemotron 3 Ultra currently shows benchmark results led by IF Bench (4 / 282, score 81.70), IMO-AnswerBench (1 / 24, score 92.30), Pinch Bench (2 / 38, score 90). This page also compares it with 3 competitor models, including performance and pricing views when available.

Benchmark Results

Nemotron 3 Ultra

Benchmark Results

General Knowledge

8 evaluations

Benchmark / mode

Score

Rank/total

GPQA

Thinking Mode

1 / 17

MMLU-Pro

Thinking Mode

86.80

17 / 176

LiveBench

Standard Mode

51.78

90 / 117

AA Intelligence Index (historical versions)

Thinking Mode

38.32

25 / 28

HLE

Thinking Mode

28.40

222 / 565

HLE

Thinking Mode

26.70

236 / 565

HLE

Thinking ModeTools

37.40

154 / 565

CritPt

Thinking Mode

3.10

111 / 201

Other

1 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Thinking Mode

86.70

137 / 463

Coding and Software Engineer

4 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBench

Thinking Mode

15 / 251

SWE-bench Verified

Thinking ModeTools

70.70

58 / 116

SWE-bench Multilingual

Thinking ModeTools

67.70

26 / 30

SciCode

Thinking Mode

44.60

93 / 131

Writing and Creative Capabilities

1 evaluations

Benchmark / mode

Score

Rank/total

Creative Writing

Standard Mode

1689.30

27 / 106

Common Sense Reasoning

1 evaluations

Benchmark / mode

Score

Rank/total

SimpleBench

Thinking Mode

41.70

62 / 93

Agent Level Benchmark

3 evaluations

Benchmark / mode

Score

Rank/total

τ²-Bench - Telecom

Thinking ModeTools

83.30

102 / 264

Terminal Bench Hard

Thinking ModeTools

36.40

67 / 244

τ³-Banking

Thinking ModeTools

22.60

85 / 167

Instruction Following

1 evaluations

Benchmark / mode

Score

Rank/total

IF Bench

Thinking Mode

81.70

4 / 282

AI Agent - Information Search

1 evaluations

Benchmark / mode

Score

Rank/total

BrowseComp

Thinking ModeToolsInternet

44.40

49 / 58

Math and Reasoning

2 evaluations

Benchmark / mode

Score

Rank/total

IMO-AnswerBench

Thinking Mode

88.60

6 / 24

IMO-AnswerBench

Thinking ModeTools

92.30

1 / 24

Long Context

2 evaluations

Benchmark / mode

Score

Rank/total

AA-LCR

Thinking Mode

79.30

54 / 171

LongBench v2

Standard Mode

61.90

5 / 14

Claw-style Agent Evaluation

2 evaluations

Benchmark / mode

Score

Rank/total

Pinch Bench

Thinking ModeTools

2 / 38

PinchBench v2

unknown

89.92

3 / 45

AI Agent - Tool Usage

2 evaluations

Benchmark / mode

Score

Rank/total

Terminal-Bench 2.1

Thinking ModeTools

56.40

122 / 196

Terminal-Bench 4.0

Thinking ModeTools

0.50

86 / 91

Productivity Knowledge

4 evaluations

Benchmark / mode

Score

Rank/total

GDPval-AA v2

Thinking ModeTools

1091

84 / 107

AA-Briefcase

Thinking ModeTools

878

71 / 85

Harvey Lab-AA

Thinking ModeTools

81.72

30 / 44

AA-AnalystAgent

Thinking ModeToolsInternet

6.25

29 / 29

Multimodal Understanding

1 evaluations

Benchmark / mode

Score

Rank/total

GDP.pdf

Thinking Mode

97 / 119

Compare with other models

Competitor Comparison

Benchmark scores for Nemotron 3 Ultra compared against top models in its class

Nemotron 3 UltraGLM-5.2 DeepSeek-V4-Pro MiniMax M3

Benchmark categories:

The chart shows each model’s highest score per benchmark within the current filter. Out-of-100 benchmarks use raw heights; out-of-range benchmarks are scaled within that benchmark while labels keep the original scores.

12 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.

Benchmark	Nemotron 3 UltraCurrent	GLM-5.2	DeepSeek-V4-Pro	MiniMax M3
AA Intelligence Index (historical versions) Historical index (not comparable) 综合评估	38.32Thinking Enabled	--	--	45.40Thinking Enabled
CritPt Score 综合评估	3.10Thinking Enabled	20.90Thinking Level · High	12.90Thinking Level · High	3.70Thinking Enabled
HLE Accuracy 综合评估	37.40Thinking Enabled ｜ Tools	54.70Thinking Enabled ｜ Tools	48.20Thinking Level · Extra High ｜ Tools	39.00Thinking Enabled
LiveBench Accuracy 综合评估	51.78Standard Mode	73.18Standard Mode	71.57Standard Mode	67.26Deep Thinking Mode
MMLU-Pro Accuracy 综合评估	86.80Thinking Enabled	--	87.50Thinking Level · High	--
GPQA Diamond Accuracy 科学与综合推理	86.70Thinking Enabled	91.86Thinking Level · High	90.50Thinking Level · High	92.90Thinking Enabled
LiveCodeBench Pass @K 编程与软件工程	89.00Thinking Enabled	--	93.50Thinking Level · High	--
SciCode Score 编程与软件工程	44.60Thinking Enabled	51.20Thinking Level · High	50.80Thinking Level · High	45.37Thinking Enabled
SWE-bench Multilingual Accuracy 编程与软件工程	67.70Thinking Enabled ｜ Tools	--	76.20Thinking Level · Extra High ｜ Tools	--
SWE-bench Verified Accuracy 编程与软件工程	70.70Thinking Enabled ｜ Tools	--	80.60Thinking Level · Extra High ｜ Tools	--
Creative Writing Elo、大模型评判两两对战写作和创作	1689.30Standard Mode	1752.80Standard Mode	1552.10Standard Mode	--
SimpleBench Score (AVG@5) 常识推理	41.70Thinking Enabled	58.80Standard Mode	50.90Standard Mode	45.80Thinking Enabled

15 additional benchmarks remain in the chart above.

Standard API Pricing: Nemotron 3 Ultra vs. Peer Models

Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.

Source: DataLearnerAI. Standard text prices shown here use the default supplier.

These models use different currencies or billing units, so the page falls back to raw price values instead of a shared bar chart.

GLM-5.2

Supplier: 智谱AI

Standard input: $1.4 / 1M tokens

Standard output: $4.4 / 1M tokens

DeepSeek-V4-Pro

Supplier: DeepSeek-AI

Standard input: $0.435 / 1M tokens

Standard output: $0.87 / 1M tokens

MiniMax M3

Supplier: MiniMaxAI

Standard input: ¥2.1 / 1M tokens

Standard output: ¥8.4 / 1M tokens

Model	Supplier	Standard input	Standard output	Base price applies to
GLM-5.2	智谱AI	$1.4 / 1M tokens	$4.4 / 1M tokens	—
DeepSeek-V4-Pro	DeepSeek-AI	$0.435 / 1M tokens	$0.87 / 1M tokens	—
MiniMax M3	MiniMaxAI	¥2.1 / 1M tokens	¥8.4 / 1M tokens	—