Nova 2 Lite

Nova 2 Lite ranks first on QVHighlights.

Amazon Released 2 Dec 2025 Family Amazon Nova 2 The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Amazon Dec 2025

Closed weights Mid-size

No Epoch Capabilities Index score yet

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.

Benchmark	Score	Rank	Source
MultiChallenge	76.6%	#2 of 36	Model card
Vectara Hallucination Leaderboard (2025 dataset)	5.10%	#12 of 104	Official leaderboard
MMLU-Pro	80.9%	#68 of 151	Model card
LMArena (Elo)	1,335	#194 of 333	Official leaderboard
IFBench	70.8%	#42 of 70	Model card
Longform Creative Writing (EQ-Bench)	38.8	#101 of 119	Official leaderboard
LongCodeBench	84%	#2 of 2	Model card

Benchmark	Score		Rank	Source
AIME 2025	91%		#45 of 143	Model card
GPQA Diamond	79.6%		#114 of 284	Model card

Benchmark	Score		Rank	Source
LiveCodeBench v5	71%		#8 of 44	Model card
SWE-bench Verified	64.5%		#97 of 150	Model card

Benchmark	Score	Rank	Source
tau2-bench	76%	#43 of 69	Model card
BFCL V4	60.3%	#17 of 24	Model card
PinchBench	63.1%	#37 of 48	Official leaderboard
Terminal-Bench	32.5%	#18 of 22	Model card
ScreenSpot	83.3%	#10 of 11	Model card
PinchBench 2.0	37.6%	#53 of 57	Official leaderboard
MCP Atlas	24.6%	#40 of 40	Model card
tau2-bench Verified	70.7%	#3 of 3	Model card

Benchmark	Score	Rank	Source
QVHighlights	77.2%	#1 of 5	Model card
RealKIE-FCC Verified	62.1%	#2 of 3	Model card
MMMU-Pro	61.8%	#54 of 72	Model card
ParseBench	40.4%	#37 of 42	Official leaderboard
OCRBench v2	56.1%	#18 of 20	Model card

Benchmark	Score	Rank	Source
Aegis 2.0	85.8%	#1 of 3	Model card
Jigsaw Toxic Comment Classification	56.5%	#2 of 3	Model card
WildGuardTest	84.7%	#4 of 5	Model card

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

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