Noometry

Model comparison

GPT-5.4 nano vs Nova 2 Lite

GPT-5.4 nano is the stronger model overall, scoring 41.9 to 39.7 on the Noometry Index.

Last verified . 18 shared benchmarks.

GPT-5.4 nano OpenAI

41.9

Rank #125 Confirmed

Nova 2 Lite Amazon

39.7

Rank #161 Confirmed

Summary

  • They share 18 benchmarks with published results for both. GPT-5.4 nano scores higher in 6 categories and Nova 2 Lite in 2 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Nova 2 Lite leads 27.5 to 23.7.
  • GPT-5.4 nano is cheaper at $0.20 / $1.25 per million input/output tokens, against $0.30 / $2.50 for Nova 2 Lite.
  • Nova 2 Lite accepts more context: 1M tokens versus 400K.

Side by side

GPT-5.4 nanoNova 2 Lite
ProviderOpenAIAmazon
Noometry Index41.939.7
Released2026-03-172025-12-01
WeightsProprietaryProprietary
Context window400K1M
Max output128K64K
Input $ / M tokens$0.20$0.30
Output $ / M tokens$1.25$2.50
Results tracked4019

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding GPT-5.4 nano leads

GPT-5.4 nano: 43.6 (#84), Nova 2 Lite: 40.7 (#134)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Coding14051385
SciCode46.9%—
WeirdML49.2%—
ALE-Bench1,005—

Agentic & Tool Use Not comparable

GPT-5.4 nano: —, Nova 2 Lite: 24.1 (#121)

BenchmarkGPT-5.4 nanoNova 2 Lite
Berkeley Function Calling Leaderboard—27.1%

Reasoning Nova 2 Lite leads

GPT-5.4 nano: 23.7 (#173), Nova 2 Lite: 27.5 (#118)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Hard Prompts13811364
ARC-AGI-25.7%—
Kagi LLM Benchmark39.7%—
ARC-AGI-151.5%—
CritPt9.3%—
Chess Puzzles30%—
Mystery Game Puzzles9%—
DTBench80.3%—
LMCA36.9%—
Epoch Capabilities Index145.81—
ForecastBench57.3—

Math GPT-5.4 nano leads

GPT-5.4 nano: 40.9 (#88), Nova 2 Lite: 37.5 (#156)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Math14061359
FrontierMath (Tiers 1-3)44.9%—
FrontierMath Tier 412.2%—
OTIS Mock AIME 2024-202587.8%—
ProofBench5%—
FrontierMath (Feb 2025 set)25.9%—
FrontierMath Tier 4 (v1)6.3%—

Knowledge Nova 2 Lite leads

GPT-5.4 nano: 41.9 (#103), Nova 2 Lite: 43.0 (#94)

BenchmarkGPT-5.4 nanoNova 2 Lite
Vectara Hallucination Rate3.1%5.1%
LMArena Expert13961358
GPQA Diamond78.5%—
SimpleQA Verified11.7%—

Multimodal Not comparable

GPT-5.4 nano: 36.7 (#78), Nova 2 Lite: —

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Vision1196—

Multilingual GPT-5.4 nano leads

GPT-5.4 nano: 48.6 (#140), Nova 2 Lite: 47.1 (#153)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Non-English13591337
LMArena Chinese13921364
LMArena French13961381
LMArena German13671343
LMArena Japanese13431271
LMArena Korean13201284
LMArena Russian13631343
LMArena Spanish13711373

Instruction Following GPT-5.4 nano leads

GPT-5.4 nano: 71.9 (#144), Nova 2 Lite: 70.5 (#161)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Instruction Following13621335

Long Context Too close to call

GPT-5.4 nano: 41.6 (#137), Nova 2 Lite: 40.6 (#150)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Longer Query13661335

Writing & Preference GPT-5.4 nano leads

GPT-5.4 nano: 55.7 (#142), Nova 2 Lite: 53.9 (#154)

BenchmarkGPT-5.4 nanoNova 2 Lite
LMArena Text13721362
LMArena Creative Writing13141291
LMArena Multi-Turn13821338

Frequently asked questions

Is GPT-5.4 nano better than Nova 2 Lite?

GPT-5.4 nano is the stronger model overall, scoring 41.9 to 39.7 on the Noometry Index.

Which is cheaper, GPT-5.4 nano or Nova 2 Lite?

GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; Nova 2 Lite lists at $0.30 and $2.50.

Is GPT-5.4 nano or Nova 2 Lite better for coding?

GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 40.7 in the Noometry coding category.

Which has the bigger context window?

Nova 2 Lite does, with 1M tokens against 400K.

How many benchmarks do GPT-5.4 nano and Nova 2 Lite share?

18 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Nova 2 Lite has 19.

Related comparisons

Go deeper