Noometry

Model comparison

Claude 3 Opus vs Phi-4

Phi-4 is the stronger model overall, scoring 31.2 to 29.5 on the Noometry Index.

Last verified . 33 shared benchmarks.

Claude 3 Opus Anthropic

29.5

Rank #310 Confirmed

Phi-4 Microsoft

31.2

Rank #279 Confirmed

Summary

  • They share 33 benchmarks with published results for both. Claude 3 Opus scores higher in 5 categories and Phi-4 in 4 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Phi-4 leads 32.6 to 24.5.
  • The biggest single-benchmark swing is MATH Level 5: 37.5% for Claude 3 Opus and 64.9% for Phi-4.
  • Phi-4 has downloadable open weights; the other is API-only.

Side by side

Claude 3 OpusPhi-4
ProviderAnthropicMicrosoft
Noometry Index29.531.2
Released2024-02-292024-12-11
WeightsProprietaryOpen
Context window—128K
Max output—4K
Input $ / M tokens—$0.07
Output $ / M tokens—$0.14
Results tracked4637

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

Category by category

Coding Phi-4 leads

Claude 3 Opus: 32.9 (#267), Phi-4: 34.4 (#239)

BenchmarkClaude 3 OpusPhi-4
BigCodeBench Instruct45.5%45.5%
LiveBench Coding38.6%30.7%
LMArena Coding12641231
BigCodeBench Complete57.4%55.4%
WeirdML19.2%—
HumanEval+77.4%—
MBPP+73.3%—

Agentic & Tool Use Claude 3 Opus leads

Claude 3 Opus: 24.6 (#116), Phi-4: 22.8 (#128)

BenchmarkClaude 3 OpusPhi-4
Berkeley Function Calling Leaderboard—28.8%
Cybench10%—
BALROG—11.6%
METR Time Horizons29.5%—

Reasoning Phi-4 leads

Claude 3 Opus: 14.6 (#324), Phi-4: 17.7 (#291)

BenchmarkClaude 3 OpusPhi-4
Chess Puzzles5%1%
LiveBench Reasoning40.6%47.8%
LMArena Hard Prompts12451220
LiveBench Data Analysis57.9%45.2%
Epoch Capabilities Index126.91130.42
LiveBench49.2%41.6%
SimpleBench23.5%—
EnigmaEval0.8%—
DTBench61.6%—
LMCA17%—
ForecastBench58.4—
WinoGrande88.5%—

Math Phi-4 leads

Claude 3 Opus: 14.8 (#299), Phi-4: 20.8 (#285)

BenchmarkClaude 3 OpusPhi-4
OTIS Mock AIME 2024-20254.7%13.8%
LiveBench Math43.6%42%
LMArena Math12731246
MATH Level 537.5%64.9%

Knowledge Phi-4 leads

Claude 3 Opus: 24.5 (#267), Phi-4: 32.6 (#209)

BenchmarkClaude 3 OpusPhi-4
GPQA Diamond47.2%56.1%
Confabulations22.7%29.4%
LMArena Expert12231203
MMLU84.6%84.8%
SimpleQA Verified12.6%—
Vectara Hallucination Rate—3.7%

Multimodal Not comparable

Claude 3 Opus: 27.1 (#116), Phi-4: —

BenchmarkClaude 3 OpusPhi-4
LMArena Vision1023—

Multilingual Claude 3 Opus leads

Claude 3 Opus: 41.4 (#207), Phi-4: 37.2 (#237)

BenchmarkClaude 3 OpusPhi-4
LMArena Non-English12581197
LMArena Chinese12481212
LMArena French12751224
LMArena German12581222
LMArena Japanese12041158
LMArena Korean11871151
LMArena Russian12801209
LMArena Spanish12461234

Instruction Following Claude 3 Opus leads

Claude 3 Opus: 64.1 (#228), Phi-4: 60.4 (#251)

BenchmarkClaude 3 OpusPhi-4
LiveBench Instruction Following63.9%58.4%
LMArena Instruction Following12481201

Long Context Claude 3 Opus leads

Claude 3 Opus: 38.2 (#202), Phi-4: 36.9 (#226)

BenchmarkClaude 3 OpusPhi-4
LMArena Longer Query12591217

Writing & Preference Claude 3 Opus leads

Claude 3 Opus: 47.2 (#213), Phi-4: 40.5 (#244)

BenchmarkClaude 3 OpusPhi-4
LMArena Text12621217
LMArena Creative Writing12351182
LMArena Multi-Turn12751206
LiveBench Language50.4%25.6%
Short-Story Creative Writing—62.6%

Frequently asked questions

Is Claude 3 Opus better than Phi-4?

Phi-4 is the stronger model overall, scoring 31.2 to 29.5 on the Noometry Index.

Is Claude 3 Opus or Phi-4 better for coding?

Phi-4 scores higher on coding benchmarks: 34.4 versus 32.9 in the Noometry coding category.

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

33 benchmarks have published results for both models. Claude 3 Opus has 46 scored results on Noometry and Phi-4 has 37.

Related comparisons

Go deeper