Noometry

Model comparison

Claude 3 Haiku vs GPT-5 Nano

GPT-5 Nano is the stronger model overall, scoring 33.5 to 25.9 on the Noometry Index.

Last verified . 27 shared benchmarks.

Claude 3 Haiku Anthropic

25.9

Rank #340 Confirmed

GPT-5 Nano OpenAI

33.5

Rank #241 Confirmed

Summary

  • They share 27 benchmarks with published results for both. Claude 3 Haiku scores higher in 1 category and GPT-5 Nano in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-5 Nano leads 29.4 to 9.8.
  • The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 95.2% for GPT-5 Nano.

Side by side

Claude 3 HaikuGPT-5 Nano
ProviderAnthropicOpenAI
Noometry Index25.933.5
Released2024-03-072025-08-07
WeightsProprietaryProprietary
Context window—400K
Max output—128K
Input $ / M tokens—$0.05
Output $ / M tokens—$0.40
Results tracked3749

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

Category by category

Coding GPT-5 Nano leads

Claude 3 Haiku: 26.4 (#325), GPT-5 Nano: 33.6 (#254)

BenchmarkClaude 3 HaikuGPT-5 Nano
WeirdML9.8%38.1%
LMArena Coding11991351
SWE-bench Verified (bash only)—34.8%
BigCodeBench Instruct39.4%—
BigCodeBench Complete50.1%—
CadEval12%—
ALE-Bench—718.67
HumanEval+68.9%—
MBPP+68.8%—

Agentic & Tool Use Not comparable

Claude 3 Haiku: —, GPT-5 Nano: 25.8 (#106)

BenchmarkClaude 3 HaikuGPT-5 Nano
Terminal-Bench—21.8%
Berkeley Function Calling Leaderboard—51.5%

Reasoning Too close to call

Claude 3 Haiku: 16.3 (#307), GPT-5 Nano: 16.3 (#306)

BenchmarkClaude 3 HaikuGPT-5 Nano
Kagi LLM Benchmark34.2%62.2%
LMArena Hard Prompts11741328
DTBench50.1%62.7%
LMCA8.8%7.9%
Epoch Capabilities Index118.35139.38
ForecastBench53.259.1
ARC-AGI-2—2.6%
ARC-AGI-1—20.7%
Chess Puzzles—27%
Mystery Game Puzzles—9%
WinoGrande74.2%—

Math GPT-5 Nano leads

Claude 3 Haiku: 9.8 (#319), GPT-5 Nano: 29.4 (#241)

BenchmarkClaude 3 HaikuGPT-5 Nano
OTIS Mock AIME 2024-20251.8%81.1%
LMArena Math11881317
MATH Level 514.9%95.2%
FrontierMath (Tiers 1-3)—20%
FrontierMath Tier 4—2.4%
ProofBench—12%
Omni-MATH—54.6%
FrontierMath (Feb 2025 set)—8.3%
FrontierMath Tier 4 (v1)—2.1%

Knowledge GPT-5 Nano leads

Claude 3 Haiku: 17.3 (#285), GPT-5 Nano: 35.9 (#178)

BenchmarkClaude 3 HaikuGPT-5 Nano
GPQA Diamond36.3%69.4%
LMArena Expert11481321
SimpleQA Verified—11.7%
MMLU-Pro—77.8%
Confabulations34.2%—
Vectara Hallucination Rate—10.5%
GPQA (HELM)—67.9%
MMLU73.8%—

Multimodal GPT-5 Nano leads

Claude 3 Haiku: 23.6 (#128), GPT-5 Nano: 31.3 (#108)

BenchmarkClaude 3 HaikuGPT-5 Nano
LMArena Vision9501159
VPCT—37.2%
ScienceQA72%—

Multilingual GPT-5 Nano leads

Claude 3 Haiku: 36.0 (#243), GPT-5 Nano: 45.3 (#172)

BenchmarkClaude 3 HaikuGPT-5 Nano
LMArena Non-English11781313
LMArena Chinese11551356
LMArena German11741327
LMArena Japanese11021226
LMArena Korean11091269
LMArena Russian12041296
LMArena Spanish11661360
LMArena French1195—

Instruction Following GPT-5 Nano leads

Claude 3 Haiku: 61.3 (#247), GPT-5 Nano: 75.0 (#79)

BenchmarkClaude 3 HaikuGPT-5 Nano
LMArena Instruction Following11731306
IFEval—93.2%

Long Context Claude 3 Haiku leads

Claude 3 Haiku: 36.1 (#237), GPT-5 Nano: 31.3 (#281)

BenchmarkClaude 3 HaikuGPT-5 Nano
LMArena Longer Query11901312
Fiction.LiveBench—44.4%

Writing & Preference GPT-5 Nano leads

Claude 3 Haiku: 29.7 (#291), GPT-5 Nano: 39.1 (#249)

BenchmarkClaude 3 HaikuGPT-5 Nano
LMArena Text11951320
LMArena Creative Writing11571249
EQ-Bench Creative Writing717705
LMArena Multi-Turn11901311
WildBench—80.6%

Frequently asked questions

Is Claude 3 Haiku better than GPT-5 Nano?

GPT-5 Nano is the stronger model overall, scoring 33.5 to 25.9 on the Noometry Index.

Is Claude 3 Haiku or GPT-5 Nano better for coding?

GPT-5 Nano scores higher on coding benchmarks: 33.6 versus 26.4 in the Noometry coding category.

How many benchmarks do Claude 3 Haiku and GPT-5 Nano share?

27 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GPT-5 Nano has 49.

Related comparisons

Go deeper