Haiku 4.5vsClaude 3.5 Haiku

Across 3 shared benchmarks, Haiku 4.5 leads overall: Haiku 4.5 wins 3, Claude 3.5 Haiku wins 0, with 0 ties and an average score difference of +11.23.

Anthropic
Haiku 4.5

Anthropic · 2025-10-15 · Multimodal model

Anthropic
Claude 3.5 Haiku

Anthropic · 2024-10-22 · Foundation model

Haiku 4.53 wins(100%)(0%)0 winsClaude 3.5 Haiku

Benchmark scores

Grouped by capability, sorted by largest gap within each. 3 shared benchmarks.

General Knowledge

Haiku 4.5 2/2
BenchmarkHaiku 4.5Claude 3.5 HaikuDiff
GPQA Diamond60.50138 / 178Normal (No Tools)41.60162 / 178+18.90
MMLU Pro7678 / 126Normal (No Tools)65101 / 126+11

Math and Reasoning

Haiku 4.5 1/1
BenchmarkHaiku 4.5Claude 3.5 HaikuDiff
FrontierMath4.1041 / 60Normal (No Tools)0.3057 / 60+3.80

Specs

FieldHaiku 4.5Claude 3.5 Haiku
PublisherAnthropicAnthropic
Release date2025-10-152024-10-22
Model typeMultimodal modelFoundation model
ArchitectureDenseDense
ParametersNot availableNot available
Context length200K200K
Max output64KNot available

Summary

  • Haiku 4.5leads in:General Knowledge (2/2), Math and Reasoning (1/1)

On average across the 3 shared benchmarks, Haiku 4.5 scores 11.23 higher.

Largest single-benchmark gap: GPQA Diamond — Haiku 4.5 60.50 vs Claude 3.5 Haiku 41.60 (+18.90).

Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.