The Model Index
Up to four models

Claude 3.5 Haiku vs Claude 3 Haiku

On the index (GPT-4 = 100), Claude 3.5 Haiku scores 109 and Claude 3 Haiku 95. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to Claude 3.5 Haiku
Presets
Claude 3.5 Haiku
10993–125
WeightsClosed
ReleasedNov 2024
Price in / out–
Context–
Core tests taken1 of 9
Claude 3 Haiku
9578–113
WeightsClosed
ReleasedMar 2024
Price in / out–
Context–
Core tests taken1 of 9

Claude 3.5 Haiku leads Claude 3 Haiku by 14 points, ahead on 1 of 1 shared core tests.

Test by testthird-party results

Claude 3.5 HaikuClaude 3 Haiku
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)MMLU (HELM runs)MMLU (HELM runs)MATHSAIME-style maths (OTIS mock)AIME-style maths (OTIS mock)MATH (level 5)MATH (level 5)CODINGCadEvalCadEvalWeirdMLWeirdMLAGENTSDTBenchDTBench
Every result as a table
TestClaude 3.5 HaikuClaude 3 Haiku
Reasoning & knowledge
GPQA Diamond (Epoch AI)38.1%36.3%
MMLU (HELM runs)74.3%73.8%
Maths
AIME-style maths (OTIS mock)4.3%1.8%
MATH (level 5)46.4%14.9%
Coding
CadEval32.0%12.0%
WeirdML30.7%9.8%
Agents
DTBench56.7%50.1%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG