Benchmark Almanac

Nova 2 Omni Preview

Nova 2 Omni Preview ranks first on 2 of the 18 benchmarks with five or more models scored.

Amazon Released December 2nd, 2025 Family Amazon Nova 2 The lab's model card Compare with others
24
Benchmarks with a score
2
Ranked first, out of five or more models
5
In the top three
Amazon Dec 2025

Nova 2 Omni Preview

Closed weights Mid-size

141.4 #166 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMMU-Pro 61.4% #66 /90
  • IFBench 68.7% #48 /80
  • BFCL V4 58.3% #21 /30

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
MultiChallenge 75.5% #3 of 42 Lab model card
MMLU-Pro 80.7% #73 of 167 Lab model card
IFBench 68.7% #48 of 80 Lab model card

Math and science

BenchmarkScoreRankSource
AIME 2025 92.1% #41 of 143 Lab model card

Agents and tools

BenchmarkScoreRankSource
tau2-bench 80% #37 of 79 Lab model card
tau2-bench Verified 75.7% #7 of 11 Official leaderboard
BFCL V4 58.3% #21 of 30 Lab model card
ScreenSpot 85.4% #9 of 11 Lab model card

Vision, audio and robotics

BenchmarkScoreRankSource
CoVoST2 40.7 #1 of 14 Lab model card
MMAU v05.15.25 77.8% #1 of 6 Official leaderboard
Common Voice 5.70% #2 of 5 Lab model card
MMAU 75.3% #9 of 15 Lab model card
QVHighlights 76.7% #3 of 5 Lab model card
Video-MME 77.9% #36 of 51 Lab model card
OCRBench v2 58.2% #16 of 22 Lab model card
MMMU-Pro 61.4% #66 of 90 Lab model card
RefCOCO / RefCOCO+ / RefCOCOg 86.3% #23 of 29 Lab model card
RealKIE-FCC Verified 59.8% #3 of 3 Lab model card
ImgEdit-Bench 4.12 #1 of 1 Lab model card
MAVERIX 66.6% #1 of 1 Lab model card
Multilingual LibriSpeech 4.70 #1 of 1 Lab model card

Safety and security

BenchmarkScoreRankSource
Jigsaw Toxic Comment Classification 57.8% #1 of 3 Lab model card
WildGuardTest 84.1% #5 of 5 Lab model card
Aegis 2.0 80.7% #3 of 3 Lab model card

Its lab's numbers put it at 141.4, number 166 of 404

The Self-Reported Index fits one score to the 12 capability results Amazon reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 138.0 to 144.9. How we compute it.

Its lab's number next to a leaderboard

Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.

Benchmark Lab says Board says Gap
tau2-bench Verified 73.5% Mean of τ2-bench Retail-Verified 78.3 / Airline-Verified 68.8 75.7% Leaderboard, tau2-bench-verified; gpt-5.1 user simulator −2.2 pts

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

Amazon Nova 2: Multimodal Reasoning and Generation Models (Technical Report and Model Card) Dec 2, 2025

Its system card

Every result in the longer documents the lab published with it. 4 of them aren't in the launch post.

Amazon Nova 2: Multimodal Reasoning and Generation Models Dec 2, 2025

Capabilities

Safety