GeekRadar / The comparison desk
COMPARE.
AI models
anthropic

Claude Haiku 4.5 (20251001)

Model evaluation

Capability & preference

Human preference · Arena

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

Claude Haiku 4.5 (20251001) 1396.5 Claude Haiku 4.5 (20251001) · 95% interval 1394.1–1398.9 · 153,897 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 Haiku 4.5 (20251001)

1,418.25 rating

Confidence interval: 1,411.94 – 1,424.56

10869 votes

Category
chinese
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,453.46 rating

Confidence interval: 1,449.52 – 1,457.41

39532 votes

Category
coding
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,370.52 rating

Confidence interval: 1,365.93 – 1,375.11

26454 votes

Category
creative_writing
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,414.25 rating

Confidence interval: 1,410.99 – 1,417.51

66213 votes

Category
english
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,376.06 rating

Confidence interval: 1,372.84 – 1,379.28

115102 votes

Category
exclude_ties
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,442.92 rating

Confidence interval: 1,437.14 – 1,448.70

14041 votes

Category
expert
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,410.66 rating

Confidence interval: 1,399.90 – 1,421.42

4456 votes

Category
french
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,373.16 rating

Confidence interval: 1,360.22 – 1,386.11

2563 votes

Category
german
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,419.98 rating

Confidence interval: 1,417.01 – 1,422.96

95033 votes

Category
hard_prompts
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,435.00 rating

Confidence interval: 1,431.09 – 1,438.90

41663 votes

Category
hard_prompts_english
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,399.47 rating

Confidence interval: 1,395.15 – 1,403.78

29616 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,369.15 rating

Confidence interval: 1,364.93 – 1,373.36

33466 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,400.23 rating

Confidence interval: 1,394.04 – 1,406.43

11534 votes

Category
industry_legal_and_government
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,406.01 rating

Confidence interval: 1,401.46 – 1,410.57

25029 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,429.42 rating

Confidence interval: 1,422.18 – 1,436.65

8153 votes

Category
industry_mathematical
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,393.81 rating

Confidence interval: 1,387.21 – 1,400.41

10388 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,439.16 rating

Confidence interval: 1,435.73 – 1,442.60

57845 votes

Category
industry_software_and_it_services
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,385.99 rating

Confidence interval: 1,382.01 – 1,389.97

37836 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,413.92 rating

Confidence interval: 1,410.29 – 1,417.55

49337 votes

Category
instruction_following
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,340.19 rating

Confidence interval: 1,324.38 – 1,356.00

1830 votes

Category
japanese
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,344.30 rating

Confidence interval: 1,330.87 – 1,357.73

2376 votes

Category
korean
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,426.44 rating

Confidence interval: 1,422.80 – 1,430.09

58500 votes

Category
longer_query
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,397.44 rating

Confidence interval: 1,390.50 – 1,404.37

8532 votes

Category
math
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,409.14 rating

Confidence interval: 1,404.65 – 1,413.62

26695 votes

Category
multi_turn
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,376.44 rating

Confidence interval: 1,373.48 – 1,379.40

87637 votes

Category
non_english
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,396.48 rating

Confidence interval: 1,394.10 – 1,398.85

153897 votes

Category
overall
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,343.15 rating

Confidence interval: 1,332.72 – 1,353.58

3583 votes

Category
polish
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,380.59 rating

Confidence interval: 1,375.48 – 1,385.70

17444 votes

Category
russian
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-11-11 · rating

Claude Haiku 4.5 (20251001)

1,422.09 rating

Confidence interval: 1,412.09 – 1,432.08

4571 votes

Category
spanish
Evaluated configuration
Claude Haiku 4.5 (20251001)
Evaluation type
human preference

Run it locally

Inference · memory estimate

Claude Haiku 4.5 (20251001)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Proprietary