GeekRadar / The comparison desk
COMPARE.
AI models
deepseek

deepseek-v4-flash-high-preview

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-flash-high-preview 1424.3 deepseek-v4-flash-high-preview · 95% interval 1420.3–1428.4 · 51,809 votes · 2026-10-15

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 · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,467.68 rating

Confidence interval: 1,456.34 – 1,479.02

2983 votes

Category
chinese
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,440.04 rating

Confidence interval: 1,433.82 – 1,446.26

15235 votes

Category
coding
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,399.00 rating

Confidence interval: 1,391.53 – 1,406.48

8664 votes

Category
creative_writing
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,434.66 rating

Confidence interval: 1,429.37 – 1,439.96

23007 votes

Category
english
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,414.08 rating

Confidence interval: 1,408.82 – 1,419.33

39443 votes

Category
exclude_ties
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,435.91 rating

Confidence interval: 1,427.13 – 1,444.69

5331 votes

Category
expert
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,436.44 rating

Confidence interval: 1,421.31 – 1,451.58

1954 votes

Category
french
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,415.08 rating

Confidence interval: 1,394.55 – 1,435.61

905 votes

Category
german
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,431.39 rating

Confidence interval: 1,426.58 – 1,436.19

34712 votes

Category
hard_prompts
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,438.33 rating

Confidence interval: 1,432.23 – 1,444.42

16089 votes

Category
hard_prompts_english
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,410.24 rating

Confidence interval: 1,403.32 – 1,417.16

10637 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,394.34 rating

Confidence interval: 1,387.51 – 1,401.18

11465 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,439.34 rating

Confidence interval: 1,429.51 – 1,449.17

4096 votes

Category
industry_legal_and_government
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,437.56 rating

Confidence interval: 1,430.24 – 1,444.88

8496 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,425.50 rating

Confidence interval: 1,413.72 – 1,437.28

2824 votes

Category
industry_mathematical
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,438.23 rating

Confidence interval: 1,427.88 – 1,448.58

3756 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,442.83 rating

Confidence interval: 1,437.32 – 1,448.34

21327 votes

Category
industry_software_and_it_services
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,408.89 rating

Confidence interval: 1,402.39 – 1,415.38

12788 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,416.94 rating

Confidence interval: 1,411.09 – 1,422.79

18090 votes

Category
instruction_following
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,405.41 rating

Confidence interval: 1,380.69 – 1,430.13

640 votes

Category
japanese
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,357.47 rating

Confidence interval: 1,336.75 – 1,378.18

949 votes

Category
korean
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,424.20 rating

Confidence interval: 1,418.65 – 1,429.75

23483 votes

Category
longer_query
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,425.19 rating

Confidence interval: 1,413.30 – 1,437.08

2624 votes

Category
math
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,431.55 rating

Confidence interval: 1,424.26 – 1,438.84

9228 votes

Category
multi_turn
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,409.74 rating

Confidence interval: 1,404.75 – 1,414.73

28800 votes

Category
non_english
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,424.34 rating

Confidence interval: 1,420.25 – 1,428.43

51809 votes

Category
overall
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,416.32 rating

Confidence interval: 1,398.08 – 1,434.55

1040 votes

Category
polish
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,425.80 rating

Confidence interval: 1,417.26 – 1,434.35

5791 votes

Category
russian
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating

deepseek-v4-flash-high-preview

1,417.73 rating

Confidence interval: 1,401.33 – 1,434.14

1475 votes

Category
spanish
Evaluated configuration
deepseek-v4-flash-high-preview
Evaluation type
human preference

Run it locally

Inference · memory estimate

deepseek-v4-flash-high-preview

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
MIT