Claude 3 HaikuvsGPT-5.6 Luna

Claude 3 Haiku
GPT-5.6 Luna
Benchmarks
Nonsense detection
BullshitBench v2
10%
40%Best
Prompt injection robustness
Gray Swan IPI · k = 1
8.3%
Prompt injection robustness
Gray Swan IPI · k = 10
38.6%
Prompt injection robustness
Gray Swan IPI · k = 15
43.9%
Agentic coding
CursorBench v3.2
61.1%
Agentic computer work
Frontier-Bench v0.1
14.3%
Agentic terminal coding
Terminal-Bench 2.1
82.5%
Science
GPQA Diamond
33.3%
Community preference (code)
Arena Elo (Code)
1523
Overview
CompanyAnthropicOpenAI
Release dateMar 4 2024Jun 26 2026
AccessProprietaryProprietary

Which is better: Claude 3 Haiku or GPT-5.6 Luna?

GPT-5.6 Luna leads Claude 3 Haiku on 1 of the 1 benchmark they both report (BullshitBench v2). Claude 3 Haiku shipped 844 days before GPT-5.6 Luna, so benchmark comparisons should account for the intervening progress.

Published specifications for these two models are limited — see each model page for the latest details.

On BullshitBench v2, GPT-5.6 Luna leads at 40% vs Claude 3 Haiku at 10%.

Frequently asked questions

Claude 3 Haiku was released by Anthropic on Mar 4 2024.

GPT-5.6 Luna was released by OpenAI on Jun 26 2026.

Other comparisons