Noometry

Model comparison

Claude 3.5 Haiku vs Phi-4

Phi-4 is the stronger model overall, scoring 31.2 to 29.2 on the Noometry Index.

Last verified . 34 shared benchmarks.

Claude 3.5 Haiku Anthropic

29.2

Rank #315 Confirmed

Phi-4 Microsoft

31.2

Rank #279 Confirmed

Summary

  • They share 34 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 5 categories and Phi-4 in 4 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Phi-4 leads 32.6 to 18.7.
  • The biggest single-benchmark swing is LiveBench Coding: 51.4% for Claude 3.5 Haiku and 30.7% for Phi-4.
  • Phi-4 has downloadable open weights; the other is API-only.

Side by side

Claude 3.5 HaikuPhi-4
ProviderAnthropicMicrosoft
Noometry Index29.231.2
Released2024-10-222024-12-11
WeightsProprietaryOpen
Context window—128K
Max output—4K
Input $ / M tokens—$0.07
Output $ / M tokens—$0.14
Results tracked4937

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

Category by category

Coding Phi-4 leads

Claude 3.5 Haiku: 32.9 (#265), Phi-4: 34.4 (#239)

BenchmarkClaude 3.5 HaikuPhi-4
BigCodeBench Instruct46.1%45.5%
LiveBench Coding51.4%30.7%
LMArena Coding12861231
BigCodeBench Complete59%55.4%
Aider Polyglot28%—
SciCode27.4%—
WeirdML30.7%—
CadEval32%—

Agentic & Tool Use Claude 3.5 Haiku leads

Claude 3.5 Haiku: 28.0 (#95), Phi-4: 22.8 (#128)

BenchmarkClaude 3.5 HaikuPhi-4
BALROG19.3%11.6%
Berkeley Function Calling Leaderboard—28.8%

Reasoning Too close to call

Claude 3.5 Haiku: 17.7 (#290), Phi-4: 17.7 (#291)

BenchmarkClaude 3.5 HaikuPhi-4
LiveBench Reasoning28.1%47.8%
LMArena Hard Prompts12511220
LiveBench Data Analysis48.5%45.2%
Epoch Capabilities Index127.15130.42
LiveBench43.5%41.6%
CritPt0%—
Chess Puzzles—1%
DTBench56.7%—

Math Phi-4 leads

Claude 3.5 Haiku: 14.7 (#300), Phi-4: 20.8 (#285)

BenchmarkClaude 3.5 HaikuPhi-4
OTIS Mock AIME 2024-20254.3%13.8%
LiveBench Math35.5%42%
LMArena Math12441246
MATH Level 546.4%64.9%
Omni-MATH22.4%—
FrontierMath (Feb 2025 set)0.3%—

Knowledge Phi-4 leads

Claude 3.5 Haiku: 18.7 (#281), Phi-4: 32.6 (#209)

BenchmarkClaude 3.5 HaikuPhi-4
GPQA Diamond38.1%56.1%
Confabulations36.7%29.4%
LMArena Expert12081203
MMLU74.3%84.8%
MMLU-Pro60.5%—
Vectara Hallucination Rate—3.7%
GPQA (HELM)36.3%—

Multimodal Not comparable

Claude 3.5 Haiku: 26.8 (#117), Phi-4: —

BenchmarkClaude 3.5 HaikuPhi-4
LMArena Vision1092—
GeoBench34%—

Multilingual Claude 3.5 Haiku leads

Claude 3.5 Haiku: 40.0 (#218), Phi-4: 37.2 (#237)

BenchmarkClaude 3.5 HaikuPhi-4
LMArena Non-English12381197
LMArena Chinese12291212
LMArena French12641224
LMArena German12371222
LMArena Japanese11751158
LMArena Korean11731151
LMArena Russian12531209
LMArena Spanish12611234

Instruction Following Claude 3.5 Haiku leads

Claude 3.5 Haiku: 62.9 (#234), Phi-4: 60.4 (#251)

BenchmarkClaude 3.5 HaikuPhi-4
LiveBench Instruction Following61.9%58.4%
LMArena Instruction Following12411201
IFEval79.2%—

Long Context Claude 3.5 Haiku leads

Claude 3.5 Haiku: 38.3 (#200), Phi-4: 36.9 (#226)

BenchmarkClaude 3.5 HaikuPhi-4
LMArena Longer Query12611217

Writing & Preference Claude 3.5 Haiku leads

Claude 3.5 Haiku: 42.7 (#234), Phi-4: 40.5 (#244)

BenchmarkClaude 3.5 HaikuPhi-4
LMArena Text12551217
LMArena Creative Writing12331182
Short-Story Creative Writing73.5%62.6%
LMArena Multi-Turn12651206
LiveBench Language35.4%25.6%
EQ-Bench Creative Writing1146—
WildBench76%—

Frequently asked questions

Is Claude 3.5 Haiku better than Phi-4?

Phi-4 is the stronger model overall, scoring 31.2 to 29.2 on the Noometry Index.

Is Claude 3.5 Haiku 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 3.5 Haiku and Phi-4 share?

34 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and Phi-4 has 37.

Related comparisons

Go deeper