Noometry

Model comparison

Claude 3 Opus vs Qwen3-4B

Qwen3-4B is the stronger model overall, scoring 31.9 to 29.5 on the Noometry Index.

Last verified . 3 shared benchmarks.

Claude 3 Opus Anthropic

29.5

Rank #310 Confirmed

Qwen3-4B Alibaba (Qwen)

31.9

Rank #264 Confirmed

Summary

  • They share 3 benchmarks with published results for both. Claude 3 Opus scores higher in 0 categories and Qwen3-4B in 4 categories; 4 gaps are clear of the uncertainty.
  • The widest gap is in math, where Qwen3-4B leads 29.7 to 14.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 4.7% for Claude 3 Opus and 52.2% for Qwen3-4B.
  • Qwen3-4B has downloadable open weights; the other is API-only.

Side by side

Claude 3 OpusQwen3-4B
ProviderAnthropicAlibaba (Qwen)
Noometry Index29.531.9
Released2024-02-292025-04-29
WeightsProprietaryOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked466

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

Category by category

Coding Not comparable

Claude 3 Opus: 32.9 (#267), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
WeirdML19.2%—
BigCodeBench Instruct45.5%—
LiveBench Coding38.6%—
LMArena Coding1264—
BigCodeBench Complete57.4%—
HumanEval+77.4%—
MBPP+73.3%—

Agentic & Tool Use Qwen3-4B leads

Claude 3 Opus: 24.6 (#116), Qwen3-4B: 27.6 (#100)

BenchmarkClaude 3 OpusQwen3-4B
Berkeley Function Calling Leaderboard—35.7%
Cybench10%—
METR Time Horizons29.5%—

Reasoning Qwen3-4B leads

Claude 3 Opus: 14.6 (#324), Qwen3-4B: 19.2 (#268)

BenchmarkClaude 3 OpusQwen3-4B
Chess Puzzles5%4%
SimpleBench23.5%—
EnigmaEval0.8%—
LiveBench Reasoning40.6%—
LMArena Hard Prompts1245—
DTBench61.6%—
LiveBench Data Analysis57.9%—
LMCA17%—
Epoch Capabilities Index126.91—
ForecastBench58.4—
LiveBench49.2%—
WinoGrande88.5%—

Math Qwen3-4B leads

Claude 3 Opus: 14.8 (#299), Qwen3-4B: 29.7 (#240)

BenchmarkClaude 3 OpusQwen3-4B
OTIS Mock AIME 2024-20254.7%52.2%
MathArena Final-Answer Competitions—38.5%
LiveBench Math43.6%—
LMArena Math1273—
MATH Level 537.5%—

Knowledge Qwen3-4B leads

Claude 3 Opus: 24.5 (#267), Qwen3-4B: 33.0 (#208)

BenchmarkClaude 3 OpusQwen3-4B
GPQA Diamond47.2%52.3%
SimpleQA Verified12.6%—
Confabulations22.7%—
Vectara Hallucination Rate—5.7%
LMArena Expert1223—
MMLU84.6%—

Multimodal Not comparable

Claude 3 Opus: 27.1 (#116), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
LMArena Vision1023—

Multilingual Not comparable

Claude 3 Opus: 41.4 (#207), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
LMArena Non-English1258—
LMArena Chinese1248—
LMArena French1275—
LMArena German1258—
LMArena Japanese1204—
LMArena Korean1187—
LMArena Russian1280—
LMArena Spanish1246—

Instruction Following Not comparable

Claude 3 Opus: 64.1 (#228), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
LiveBench Instruction Following63.9%—
LMArena Instruction Following1248—

Long Context Not comparable

Claude 3 Opus: 38.2 (#202), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
LMArena Longer Query1259—

Writing & Preference Not comparable

Claude 3 Opus: 47.2 (#213), Qwen3-4B: —

BenchmarkClaude 3 OpusQwen3-4B
LMArena Text1262—
LMArena Creative Writing1235—
LMArena Multi-Turn1275—
LiveBench Language50.4%—

Frequently asked questions

Is Claude 3 Opus better than Qwen3-4B?

Qwen3-4B is the stronger model overall, scoring 31.9 to 29.5 on the Noometry Index.

How many benchmarks do Claude 3 Opus and Qwen3-4B share?

3 benchmarks have published results for both models. Claude 3 Opus has 46 scored results on Noometry and Qwen3-4B has 6.

Related comparisons

Go deeper