GeekRadar / The comparison desk
COMPARE.
AI models
anthropic

Claude Sonnet 4 (Thinking 32K, 20250514)

Model evaluation

Capability & preference

Human preference · Arena

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

Claude Sonnet 4 (Thinking 32K, 20250514) 1350.9 Claude Sonnet 4 (Thinking 32K, 20250514) · 95% interval 1346.4–1355.4 · 34,128 votes · 2026-10-28

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 · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,349.90 rating

Confidence interval: 1,337.55 – 1,362.25

2376 votes

Category
chinese
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,414.17 rating

Confidence interval: 1,406.28 – 1,422.07

6343 votes

Category
coding
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,345.10 rating

Confidence interval: 1,335.62 – 1,354.58

4249 votes

Category
creative_writing
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,364.49 rating

Confidence interval: 1,358.68 – 1,370.29

14846 votes

Category
english
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,311.09 rating

Confidence interval: 1,305.21 – 1,316.97

25035 votes

Category
exclude_ties
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,372.20 rating

Confidence interval: 1,357.72 – 1,386.68

1701 votes

Category
expert
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,356.38 rating

Confidence interval: 1,329.61 – 1,383.16

473 votes

Category
french
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,317.62 rating

Confidence interval: 1,296.80 – 1,338.44

865 votes

Category
german
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,372.41 rating

Confidence interval: 1,366.68 – 1,378.13

15419 votes

Category
hard_prompts
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,385.21 rating

Confidence interval: 1,377.58 – 1,392.83

6982 votes

Category
hard_prompts_english
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,314.26 rating

Confidence interval: 1,305.89 – 1,322.63

5776 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,332.02 rating

Confidence interval: 1,323.98 – 1,340.06

6046 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,336.68 rating

Confidence interval: 1,323.51 – 1,349.85

2206 votes

Category
industry_legal_and_government
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,345.49 rating

Confidence interval: 1,337.15 – 1,353.84

5616 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,379.18 rating

Confidence interval: 1,365.83 – 1,392.53

1914 votes

Category
industry_mathematical
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,326.00 rating

Confidence interval: 1,312.63 – 1,339.37

2125 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,383.76 rating

Confidence interval: 1,377.41 – 1,390.12

11254 votes

Category
industry_software_and_it_services
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,356.91 rating

Confidence interval: 1,349.50 – 1,364.31

7375 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,375.99 rating

Confidence interval: 1,368.97 – 1,383.00

8526 votes

Category
instruction_following
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,288.03 rating

Confidence interval: 1,265.78 – 1,310.29

824 votes

Category
japanese
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,271.13 rating

Confidence interval: 1,248.22 – 1,294.05

729 votes

Category
korean
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,397.59 rating

Confidence interval: 1,389.91 – 1,405.27

6725 votes

Category
longer_query
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,374.88 rating

Confidence interval: 1,362.01 – 1,387.76

2013 votes

Category
math
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,376.27 rating

Confidence interval: 1,367.91 – 1,384.63

5635 votes

Category
multi_turn
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,332.59 rating

Confidence interval: 1,327.21 – 1,337.96

19277 votes

Category
non_english
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,350.94 rating

Confidence interval: 1,346.44 – 1,355.44

34128 votes

Category
overall
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,324.41 rating

Confidence interval: 1,313.05 – 1,335.78

3271 votes

Category
polish
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,355.17 rating

Confidence interval: 1,343.32 – 1,367.03

2410 votes

Category
russian
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

Claude Sonnet 4 (Thinking 32K, 20250514)

1,356.76 rating

Confidence interval: 1,335.42 – 1,378.09

835 votes

Category
spanish
Evaluated configuration
Claude Sonnet 4 (Thinking 32K, 20250514)
Evaluation type
human preference

Run it locally

Inference · memory estimate

Claude Sonnet 4 (Thinking 32K, 20250514)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
Proprietary