GeekRadar / The comparison desk
COMPARE.
AI models
anthropic

Claude 3.5 Haiku (20241022)

Model evaluation

Capability & preference

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Claude 3.5 Haiku (20241022) 1255.4 Claude 3.5 Haiku (20241022) · 95% interval 1252.2–1258.6 · 69,006 votes · 2026-11-11

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Arena text preference / chinesehuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,228.87 rating

Confidence interval: 1,219.54 – 1,238.20

4655 votes

Category
chinese
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,286.23 rating

Confidence interval: 1,280.26 – 1,292.19

11135 votes

Category
coding
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,232.37 rating

Confidence interval: 1,225.93 – 1,238.82

9834 votes

Category
creative_writing
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,267.47 rating

Confidence interval: 1,263.41 – 1,271.53

35484 votes

Category
english
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,171.09 rating

Confidence interval: 1,166.42 – 1,175.76

48565 votes

Category
exclude_ties
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,207.16 rating

Confidence interval: 1,196.63 – 1,217.70

3517 votes

Category
expert
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,263.81 rating

Confidence interval: 1,241.85 – 1,285.78

837 votes

Category
french
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,236.50 rating

Confidence interval: 1,221.42 – 1,251.57

1737 votes

Category
german
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,250.83 rating

Confidence interval: 1,246.24 – 1,255.43

23525 votes

Category
hard_prompts
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,264.19 rating

Confidence interval: 1,258.36 – 1,270.03

12319 votes

Category
hard_prompts_english
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,222.51 rating

Confidence interval: 1,215.66 – 1,229.36

9066 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,223.83 rating

Confidence interval: 1,218.02 – 1,229.63

12371 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,258.78 rating

Confidence interval: 1,248.99 – 1,268.58

4228 votes

Category
industry_legal_and_government
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,247.21 rating

Confidence interval: 1,241.03 – 1,253.39

11836 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,248.94 rating

Confidence interval: 1,240.94 – 1,256.93

5752 votes

Category
industry_mathematical
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,227.51 rating

Confidence interval: 1,216.62 – 1,238.40

3637 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,272.25 rating

Confidence interval: 1,267.32 – 1,277.18

19585 votes

Category
industry_software_and_it_services
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,256.29 rating

Confidence interval: 1,251.14 – 1,261.44

16804 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,240.47 rating

Confidence interval: 1,235.81 – 1,245.13

21913 votes

Category
instruction_following
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,175.29 rating

Confidence interval: 1,158.93 – 1,191.66

1546 votes

Category
japanese
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,173.34 rating

Confidence interval: 1,154.68 – 1,192.00

1210 votes

Category
korean
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,260.42 rating

Confidence interval: 1,254.36 – 1,266.49

10954 votes

Category
longer_query
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,243.74 rating

Confidence interval: 1,236.22 – 1,251.27

6328 votes

Category
math
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,265.00 rating

Confidence interval: 1,258.78 – 1,271.21

11168 votes

Category
multi_turn
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,237.86 rating

Confidence interval: 1,233.76 – 1,241.95

33514 votes

Category
non_english
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,255.40 rating

Confidence interval: 1,252.17 – 1,258.63

69006 votes

Category
overall
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,249.33 rating

Confidence interval: 1,236.62 – 1,262.05

2715 votes

Category
polish
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,252.59 rating

Confidence interval: 1,245.13 – 1,260.05

6642 votes

Category
russian
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-11-11 · rating

Claude 3.5 Haiku (20241022)

1,261.34 rating

Confidence interval: 1,241.34 – 1,281.34

961 votes

Category
spanish
Evaluated configuration
Claude 3.5 Haiku (20241022)
Evaluation type
human preference

Run it locally

Inference · memory estimate

Claude 3.5 Haiku (20241022)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Proprietary