Noometry

Model comparison

Claude 3 Haiku vs GPT-6 Luna

GPT-6 Luna is the stronger model overall, scoring 53.3 to 25.9 on the Noometry Index.

Last verified . 23 shared benchmarks.

Claude 3 Haiku Anthropic

25.9

Rank #340 Confirmed

GPT-6 Luna OpenAI

53.3

Rank #36 Confirmed

Summary

  • They share 23 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and GPT-6 Luna in 9 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-6 Luna leads 76.1 to 9.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.8% for Claude 3 Haiku and 98.9% for GPT-6 Luna.

Side by side

Claude 3 HaikuGPT-6 Luna
ProviderAnthropicOpenAI
Noometry Index25.953.3
Released2024-03-072026-09-22
WeightsProprietaryProprietary
Context window—1.05M
Max output—128K
Input $ / M tokens—$0.10
Output $ / M tokens—$0.50
Results tracked3742

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

Category by category

Coding GPT-6 Luna leads

Claude 3 Haiku: 26.4 (#325), GPT-6 Luna: 55.5 (#25)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Coding11991439
DeepSWE—66.6%
FrontierCode—42.4%
LMArena WebDev—1581
SciCode—54.6%
WeirdML9.8%—
BigCodeBench Instruct39.4%—
BigCodeBench Complete50.1%—
CadEval12%—
ALE-Bench—1,577
HumanEval+68.9%—
MBPP+68.8%—

Agentic & Tool Use Not comparable

Claude 3 Haiku: —, GPT-6 Luna: 33.3 (#54)

BenchmarkClaude 3 HaikuGPT-6 Luna
APEX-Agents—44.3%
GDP.pdf—23%

Reasoning GPT-6 Luna leads

Claude 3 Haiku: 16.3 (#307), GPT-6 Luna: 48.2 (#41)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Hard Prompts11741411
DTBench50.1%90.1%
LMCA8.8%44.5%
Epoch Capabilities Index118.35156.28
ARC-AGI-2—59.3%
Kagi LLM Benchmark34.2%—
NYT Connections (extended)—68.7%
ARC-AGI-1—86.7%
CritPt—19.4%
Chess Puzzles—31%
Mystery Game Puzzles—7%
ForecastBench53.2—
WinoGrande74.2%—

Math GPT-6 Luna leads

Claude 3 Haiku: 9.8 (#319), GPT-6 Luna: 76.1 (#15)

BenchmarkClaude 3 HaikuGPT-6 Luna
OTIS Mock AIME 2024-20251.8%98.9%
LMArena Math11881416
FrontierMath (Tiers 1-3)—78.9%
FrontierMath Tier 4—56.1%
ProofBench—64%
MATH Level 514.9%—

Knowledge GPT-6 Luna leads

Claude 3 Haiku: 17.3 (#285), GPT-6 Luna: 57.0 (#41)

BenchmarkClaude 3 HaikuGPT-6 Luna
GPQA Diamond36.3%90.5%
LMArena Expert11481444
SimpleQA Verified—41.4%
Confabulations34.2%—
MMLU73.8%—

Multimodal GPT-6 Luna leads

Claude 3 Haiku: 23.6 (#128), GPT-6 Luna: 42.4 (#30)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Vision9501217
Blueprint-Bench 2—31.2%
Furniture Assembly—44.2%
ScienceQA72%—

Multilingual GPT-6 Luna leads

Claude 3 Haiku: 36.0 (#243), GPT-6 Luna: 50.5 (#117)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Non-English11781386
LMArena Chinese11551433
LMArena French11951420
LMArena German11741369
LMArena Japanese11021369
LMArena Korean11091360
LMArena Russian12041394
LMArena Spanish11661393

Instruction Following GPT-6 Luna leads

Claude 3 Haiku: 61.3 (#247), GPT-6 Luna: 74.3 (#99)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Instruction Following11731409

Long Context GPT-6 Luna leads

Claude 3 Haiku: 36.1 (#237), GPT-6 Luna: 43.0 (#111)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Longer Query11901409

Writing & Preference GPT-6 Luna leads

Claude 3 Haiku: 29.7 (#291), GPT-6 Luna: 58.3 (#119)

BenchmarkClaude 3 HaikuGPT-6 Luna
LMArena Text11951391
LMArena Creative Writing11571363
LMArena Multi-Turn11901396
EQ-Bench Creative Writing717—

Frequently asked questions

Is Claude 3 Haiku better than GPT-6 Luna?

GPT-6 Luna is the stronger model overall, scoring 53.3 to 25.9 on the Noometry Index.

Is Claude 3 Haiku or GPT-6 Luna better for coding?

GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 26.4 in the Noometry coding category.

How many benchmarks do Claude 3 Haiku and GPT-6 Luna share?

23 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GPT-6 Luna has 42.

Related comparisons

Go deeper