Noometry

Model comparison

Claude 3 Haiku vs GPT-5.4 nano

GPT-5.4 nano is the stronger model overall, scoring 41.9 to 25.9 on the Noometry Index.

Last verified . 26 shared benchmarks.

Claude 3 Haiku Anthropic

25.9

Rank #340 Confirmed

GPT-5.4 nano OpenAI

41.9

Rank #125 Confirmed

Summary

  • They share 26 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and GPT-5.4 nano in 9 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-5.4 nano leads 40.9 to 9.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.8% for Claude 3 Haiku and 87.8% for GPT-5.4 nano.

Side by side

Claude 3 HaikuGPT-5.4 nano
ProviderAnthropicOpenAI
Noometry Index25.941.9
Released2024-03-072026-03-17
WeightsProprietaryProprietary
Context window—400K
Max output—128K
Input $ / M tokens—$0.20
Output $ / M tokens—$1.25
Results tracked3740

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

Category by category

Coding GPT-5.4 nano leads

Claude 3 Haiku: 26.4 (#325), GPT-5.4 nano: 43.6 (#84)

BenchmarkClaude 3 HaikuGPT-5.4 nano
WeirdML9.8%49.2%
LMArena Coding11991405
SciCode—46.9%
BigCodeBench Instruct39.4%—
BigCodeBench Complete50.1%—
CadEval12%—
ALE-Bench—1,005
HumanEval+68.9%—
MBPP+68.8%—

Reasoning GPT-5.4 nano leads

Claude 3 Haiku: 16.3 (#307), GPT-5.4 nano: 23.7 (#173)

BenchmarkClaude 3 HaikuGPT-5.4 nano
Kagi LLM Benchmark34.2%39.7%
LMArena Hard Prompts11741381
DTBench50.1%80.3%
LMCA8.8%36.9%
Epoch Capabilities Index118.35145.81
ForecastBench53.257.3
ARC-AGI-2—5.7%
ARC-AGI-1—51.5%
CritPt—9.3%
Chess Puzzles—30%
Mystery Game Puzzles—9%
WinoGrande74.2%—

Math GPT-5.4 nano leads

Claude 3 Haiku: 9.8 (#319), GPT-5.4 nano: 40.9 (#88)

BenchmarkClaude 3 HaikuGPT-5.4 nano
OTIS Mock AIME 2024-20251.8%87.8%
LMArena Math11881406
FrontierMath (Tiers 1-3)—44.9%
FrontierMath Tier 4—12.2%
ProofBench—5%
MATH Level 514.9%—
FrontierMath (Feb 2025 set)—25.9%
FrontierMath Tier 4 (v1)—6.3%

Knowledge GPT-5.4 nano leads

Claude 3 Haiku: 17.3 (#285), GPT-5.4 nano: 41.9 (#103)

BenchmarkClaude 3 HaikuGPT-5.4 nano
GPQA Diamond36.3%78.5%
LMArena Expert11481396
SimpleQA Verified—11.7%
Confabulations34.2%—
Vectara Hallucination Rate—3.1%
MMLU73.8%—

Multimodal GPT-5.4 nano leads

Claude 3 Haiku: 23.6 (#128), GPT-5.4 nano: 36.7 (#78)

BenchmarkClaude 3 HaikuGPT-5.4 nano
LMArena Vision9501196
ScienceQA72%—

Multilingual GPT-5.4 nano leads

Claude 3 Haiku: 36.0 (#243), GPT-5.4 nano: 48.6 (#140)

BenchmarkClaude 3 HaikuGPT-5.4 nano
LMArena Non-English11781359
LMArena Chinese11551392
LMArena French11951396
LMArena German11741367
LMArena Japanese11021343
LMArena Korean11091320
LMArena Russian12041363
LMArena Spanish11661371

Instruction Following GPT-5.4 nano leads

Claude 3 Haiku: 61.3 (#247), GPT-5.4 nano: 71.9 (#144)

BenchmarkClaude 3 HaikuGPT-5.4 nano
LMArena Instruction Following11731362

Long Context GPT-5.4 nano leads

Claude 3 Haiku: 36.1 (#237), GPT-5.4 nano: 41.6 (#137)

BenchmarkClaude 3 HaikuGPT-5.4 nano
LMArena Longer Query11901366

Writing & Preference GPT-5.4 nano leads

Claude 3 Haiku: 29.7 (#291), GPT-5.4 nano: 55.7 (#142)

BenchmarkClaude 3 HaikuGPT-5.4 nano
LMArena Text11951372
LMArena Creative Writing11571314
LMArena Multi-Turn11901382
EQ-Bench Creative Writing717—

Frequently asked questions

Is Claude 3 Haiku better than GPT-5.4 nano?

GPT-5.4 nano is the stronger model overall, scoring 41.9 to 25.9 on the Noometry Index.

Is Claude 3 Haiku or GPT-5.4 nano better for coding?

GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 26.4 in the Noometry coding category.

How many benchmarks do Claude 3 Haiku and GPT-5.4 nano share?

26 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GPT-5.4 nano has 40.

Related comparisons

Go deeper