GeekRadar / The comparison desk
COMPARE.
AI models
deepseek

DeepSeek-R1 (0528)

Model evaluation

Capability & preference

Human preference · Arena

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

DeepSeek-R1 (0528) 1427.7 DeepSeek-R1 (0528) · 95% interval 1422.1–1433.4 · 18,165 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

DeepSeek-R1 (0528)

1,441.92 rating

Confidence interval: 1,424.30 – 1,459.55

1186 votes

Category
chinese
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,426.81 rating

Confidence interval: 1,415.40 – 1,438.21

2690 votes

Category
coding
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,405.25 rating

Confidence interval: 1,392.83 – 1,417.66

2408 votes

Category
creative_writing
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,437.20 rating

Confidence interval: 1,429.85 – 1,444.55

8162 votes

Category
english
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,419.80 rating

Confidence interval: 1,412.00 – 1,427.61

13077 votes

Category
exclude_ties
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,393.29 rating

Confidence interval: 1,373.65 – 1,412.94

887 votes

Category
expert
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,417.50 rating

Confidence interval: 1,379.95 – 1,455.05

237 votes

Category
french
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,404.25 rating

Confidence interval: 1,379.46 – 1,429.05

597 votes

Category
german
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,415.30 rating

Confidence interval: 1,407.62 – 1,422.99

6881 votes

Category
hard_prompts
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,421.76 rating

Confidence interval: 1,410.99 – 1,432.53

3106 votes

Category
hard_prompts_english
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,401.66 rating

Confidence interval: 1,390.47 – 1,412.85

2851 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,400.25 rating

Confidence interval: 1,389.61 – 1,410.88

3183 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,418.66 rating

Confidence interval: 1,402.42 – 1,434.90

1370 votes

Category
industry_legal_and_government
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,430.36 rating

Confidence interval: 1,419.79 – 1,440.93

3386 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,415.17 rating

Confidence interval: 1,395.86 – 1,434.48

919 votes

Category
industry_mathematical
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,444.84 rating

Confidence interval: 1,427.33 – 1,462.36

1229 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,434.72 rating

Confidence interval: 1,426.37 – 1,443.07

5530 votes

Category
industry_software_and_it_services
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,401.47 rating

Confidence interval: 1,391.97 – 1,410.96

4199 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,381.54 rating

Confidence interval: 1,372.01 – 1,391.07

4015 votes

Category
instruction_following
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,391.08 rating

Confidence interval: 1,366.22 – 1,415.95

620 votes

Category
japanese
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,358.92 rating

Confidence interval: 1,324.55 – 1,393.29

324 votes

Category
korean
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,391.08 rating

Confidence interval: 1,380.50 – 1,401.67

3130 votes

Category
longer_query
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,399.98 rating

Confidence interval: 1,380.05 – 1,419.92

861 votes

Category
math
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,404.40 rating

Confidence interval: 1,393.43 – 1,415.37

2847 votes

Category
multi_turn
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,411.71 rating

Confidence interval: 1,404.88 – 1,418.54

9997 votes

Category
non_english
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,427.71 rating

Confidence interval: 1,422.06 – 1,433.37

18165 votes

Category
overall
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,428.11 rating

Confidence interval: 1,413.96 – 1,442.27

1980 votes

Category
polish
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,422.77 rating

Confidence interval: 1,408.11 – 1,437.42

1523 votes

Category
russian
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

DeepSeek-R1 (0528)

1,411.43 rating

Confidence interval: 1,375.94 – 1,446.93

261 votes

Category
spanish
Evaluated configuration
DeepSeek-R1 (0528)
Evaluation type
human preference

Run it locally

Inference · memory estimate

DeepSeek-R1 (0528)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
MIT