Noometry

Model comparison

Amazon Nova Pro vs Qwen3-4B

Amazon Nova Pro and Qwen3-4B score almost the same on the Noometry Index (31.0 vs 31.9), so choose on price, context window or the category you care about most.

Last verified . 2 shared benchmarks.

Amazon Nova Pro Amazon

31.0

Rank #281 Confirmed

Qwen3-4B Alibaba (Qwen)

31.9

Rank #264 Confirmed

Summary

  • They share 2 benchmarks with published results for both. Amazon Nova Pro scores higher in 1 category and Qwen3-4B in 3 categories; 3 gaps are clear of the uncertainty.
  • The widest gap is in agentic & tool use, where Qwen3-4B leads 27.6 to 16.7.
  • The biggest single-benchmark swing is Berkeley Function Calling Leaderboard: 25% for Amazon Nova Pro and 35.7% for Qwen3-4B.
  • Qwen3-4B has downloadable open weights; the other is API-only.

Side by side

Amazon Nova ProQwen3-4B
ProviderAmazonAlibaba (Qwen)
Noometry Index31.031.9
Released2024-12-032025-04-29
WeightsProprietaryOpen
Context window300K—
Max output10K—
Input $ / M tokens$0.80—
Output $ / M tokens$3.20—
Results tracked386

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

Category by category

Coding Not comparable

Amazon Nova Pro: 35.1 (#229), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LiveBench Coding38.1%—
LMArena Coding1270—

Agentic & Tool Use Qwen3-4B leads

Amazon Nova Pro: 16.7 (#147), Qwen3-4B: 27.6 (#100)

BenchmarkAmazon Nova ProQwen3-4B
Berkeley Function Calling Leaderboard25%35.7%
TheAgentCompany1.7%—

Reasoning Too close to call

Amazon Nova Pro: 20.0 (#243), Qwen3-4B: 19.2 (#268)

BenchmarkAmazon Nova ProQwen3-4B
Chess Puzzles—4%
LiveBench Reasoning32.6%—
LMArena Hard Prompts1246—
LiveBench Data Analysis48.3%—
Epoch Capabilities Index123.8—
LiveBench43.5%—

Math Qwen3-4B leads

Amazon Nova Pro: 28.5 (#243), Qwen3-4B: 29.7 (#240)

BenchmarkAmazon Nova ProQwen3-4B
MathArena Final-Answer Competitions—38.5%
OTIS Mock AIME 2024-2025—52.2%
Omni-MATH24.2%—
LiveBench Math38%—
LMArena Math1252—

Knowledge Qwen3-4B leads

Amazon Nova Pro: 27.4 (#250), Qwen3-4B: 33.0 (#208)

BenchmarkAmazon Nova ProQwen3-4B
Vectara Hallucination Rate5.1%5.7%
GPQA Diamond—52.3%
Humanity's Last Exam4.4%—
MMLU-Pro67.3%—
Confabulations30.1%—
GPQA (HELM)44.6%—
LMArena Expert1211—
MMLU82%—

Multimodal Not comparable

Amazon Nova Pro: 25.0 (#126), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LMArena Vision980—

Multilingual Not comparable

Amazon Nova Pro: 39.7 (#223), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LMArena Non-English1234—
LMArena Chinese1244—
LMArena French1271—
LMArena German1243—
LMArena Japanese1200—
LMArena Korean1203—
LMArena Russian1240—
LMArena Spanish1182—

Instruction Following Not comparable

Amazon Nova Pro: 64.9 (#226), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LiveBench Instruction Following67.1%—
IFEval81.5%—
LMArena Instruction Following1235—

Long Context Not comparable

Amazon Nova Pro: 38.1 (#205), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LMArena Longer Query1255—

Writing & Preference Not comparable

Amazon Nova Pro: 43.9 (#226), Qwen3-4B: —

BenchmarkAmazon Nova ProQwen3-4B
LMArena Text1259—
LMArena Creative Writing1212—
Short-Story Creative Writing60.5%—
WildBench77.7%—
LMArena Multi-Turn1246—
LiveBench Language37%—

Frequently asked questions

Is Amazon Nova Pro better than Qwen3-4B?

Amazon Nova Pro and Qwen3-4B score almost the same on the Noometry Index (31.0 vs 31.9), so choose on price, context window or the category you care about most.

How many benchmarks do Amazon Nova Pro and Qwen3-4B share?

2 benchmarks have published results for both models. Amazon Nova Pro has 38 scored results on Noometry and Qwen3-4B has 6.

Related comparisons

Go deeper