Coverage (30d)
5vs3
This Week
2vs1
Evidence
1 articlesRelationships
1Timeline
Claude Haiku 4.52026-04-05
Achieved highest adversarial cost ($10.21) in ACE benchmark, 10x more secure than next model
Claude Haiku 4.52026-03-13
Became preferred AI model for GitHub Copilot in Visual Studio 2026 among developers
Ecosystem
Claude Haiku 4.5
competes withGPT-4o Nano1 src
GPT-4o Nano
No mapped relationships
Benchmarks
mmlu pro
Claude Haiku 4.578
GPT-4o Nano—
swe bench verified
Claude Haiku 4.552.6
GPT-4o Nano—