GeekRadar / The comparison desk
COMPARE.
AI models
deepseek

DeepSeek-V4-Pro (High, 20260813)

Model evaluation

Capability & preference

Human preference · Arena

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

DeepSeek-V4-Pro (High, 20260813) 1446.0 DeepSeek-V4-Pro (High, 20260813) · 95% interval 1439.5–1452.6 · 9,787 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-V4-Pro (High, 20260813)

1,482.46 rating

Confidence interval: 1,460.07 – 1,504.85

707 votes

Category
chinese
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,468.64 rating

Confidence interval: 1,457.33 – 1,479.96

2809 votes

Category
coding
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,439.55 rating

Confidence interval: 1,425.67 – 1,453.44

1976 votes

Category
creative_writing
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,458.13 rating

Confidence interval: 1,448.40 – 1,467.87

3761 votes

Category
english
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,442.76 rating

Confidence interval: 1,434.08 – 1,451.43

7161 votes

Category
exclude_ties
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,459.32 rating

Confidence interval: 1,440.95 – 1,477.69

1032 votes

Category
expert
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,446.67 rating

Confidence interval: 1,413.64 – 1,479.69

316 votes

Category
french
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,457.85 rating

Confidence interval: 1,450.14 – 1,465.56

6522 votes

Category
hard_prompts
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,463.44 rating

Confidence interval: 1,451.42 – 1,475.46

2490 votes

Category
hard_prompts_english
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,442.36 rating

Confidence interval: 1,428.43 – 1,456.28

1810 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,424.04 rating

Confidence interval: 1,411.59 – 1,436.48

2447 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,459.31 rating

Confidence interval: 1,439.08 – 1,479.55

825 votes

Category
industry_legal_and_government
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,458.42 rating

Confidence interval: 1,443.32 – 1,473.51

1525 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,443.21 rating

Confidence interval: 1,417.23 – 1,469.19

516 votes

Category
industry_mathematical
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,446.46 rating

Confidence interval: 1,423.43 – 1,469.50

626 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,461.48 rating

Confidence interval: 1,452.07 – 1,470.90

4093 votes

Category
industry_software_and_it_services
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,443.84 rating

Confidence interval: 1,431.87 – 1,455.80

2615 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,448.64 rating

Confidence interval: 1,438.38 – 1,458.90

3474 votes

Category
instruction_following
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,447.16 rating

Confidence interval: 1,406.58 – 1,487.73

200 votes

Category
korean
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,457.61 rating

Confidence interval: 1,448.39 – 1,466.83

4558 votes

Category
longer_query
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,441.89 rating

Confidence interval: 1,414.12 – 1,469.65

407 votes

Category
math
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,453.98 rating

Confidence interval: 1,438.62 – 1,469.34

1564 votes

Category
multi_turn
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,430.81 rating

Confidence interval: 1,422.91 – 1,438.72

6026 votes

Category
non_english
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,446.04 rating

Confidence interval: 1,439.47 – 1,452.61

9787 votes

Category
overall
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,447.35 rating

Confidence interval: 1,430.66 – 1,464.04

1219 votes

Category
russian
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

DeepSeek-V4-Pro (High, 20260813)

1,435.26 rating

Confidence interval: 1,403.06 – 1,467.46

299 votes

Category
spanish
Evaluated configuration
DeepSeek-V4-Pro (High, 20260813)
Evaluation type
human preference

Run it locally

Inference · memory estimate

DeepSeek-V4-Pro (High, 20260813)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
MIT