GeekRadar / The comparison desk
COMPARE.
AI models
deepseek

DeepSeek-V4-Flash Preview (High)

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 Preview (High) 1424.3 DeepSeek-V4-Flash Preview (High) · 95% interval 1420.2–1428.4 · 51,763 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-Flash Preview (High)

1,467.39 rating

Confidence interval: 1,456.03 – 1,478.74

2971 votes

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

DeepSeek-V4-Flash Preview (High)

1,440.17 rating

Confidence interval: 1,433.95 – 1,446.39

15235 votes

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

DeepSeek-V4-Flash Preview (High)

1,398.48 rating

Confidence interval: 1,391.01 – 1,405.95

8642 votes

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

DeepSeek-V4-Flash Preview (High)

1,434.69 rating

Confidence interval: 1,429.40 – 1,439.98

22996 votes

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

DeepSeek-V4-Flash Preview (High)

1,414.01 rating

Confidence interval: 1,408.76 – 1,419.26

39392 votes

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

DeepSeek-V4-Flash Preview (High)

1,436.34 rating

Confidence interval: 1,427.58 – 1,445.11

5331 votes

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

DeepSeek-V4-Flash Preview (High)

1,435.84 rating

Confidence interval: 1,420.72 – 1,450.97

1949 votes

Category
french
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,415.35 rating

Confidence interval: 1,394.87 – 1,435.84

902 votes

Category
german
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,431.54 rating

Confidence interval: 1,426.73 – 1,436.34

34668 votes

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

DeepSeek-V4-Flash Preview (High)

1,438.56 rating

Confidence interval: 1,432.47 – 1,444.65

16078 votes

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

DeepSeek-V4-Flash Preview (High)

1,410.43 rating

Confidence interval: 1,403.52 – 1,417.35

10631 votes

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

DeepSeek-V4-Flash Preview (High)

1,393.62 rating

Confidence interval: 1,386.79 – 1,400.46

11451 votes

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

DeepSeek-V4-Flash Preview (High)

1,439.18 rating

Confidence interval: 1,429.35 – 1,449.01

4089 votes

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

DeepSeek-V4-Flash Preview (High)

1,437.11 rating

Confidence interval: 1,429.79 – 1,444.42

8482 votes

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

DeepSeek-V4-Flash Preview (High)

1,426.29 rating

Confidence interval: 1,414.53 – 1,438.05

2827 votes

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

DeepSeek-V4-Flash Preview (High)

1,437.23 rating

Confidence interval: 1,426.88 – 1,447.59

3743 votes

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

DeepSeek-V4-Flash Preview (High)

1,442.71 rating

Confidence interval: 1,437.21 – 1,448.21

21326 votes

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

DeepSeek-V4-Flash Preview (High)

1,408.48 rating

Confidence interval: 1,401.99 – 1,414.97

12770 votes

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

DeepSeek-V4-Flash Preview (High)

1,417.14 rating

Confidence interval: 1,411.29 – 1,422.98

18064 votes

Category
instruction_following
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,404.85 rating

Confidence interval: 1,380.25 – 1,429.46

641 votes

Category
japanese
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,358.29 rating

Confidence interval: 1,337.59 – 1,378.99

943 votes

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

DeepSeek-V4-Flash Preview (High)

1,423.92 rating

Confidence interval: 1,418.37 – 1,429.46

23474 votes

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

DeepSeek-V4-Flash Preview (High)

1,425.36 rating

Confidence interval: 1,413.48 – 1,437.23

2625 votes

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

DeepSeek-V4-Flash Preview (High)

1,431.58 rating

Confidence interval: 1,424.30 – 1,438.87

9226 votes

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

DeepSeek-V4-Flash Preview (High)

1,409.52 rating

Confidence interval: 1,404.53 – 1,414.51

28765 votes

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

DeepSeek-V4-Flash Preview (High)

1,424.31 rating

Confidence interval: 1,420.22 – 1,428.40

51763 votes

Category
overall
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,418.01 rating

Confidence interval: 1,399.72 – 1,436.31

1030 votes

Category
polish
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

DeepSeek-V4-Flash Preview (High)

1,425.01 rating

Confidence interval: 1,416.47 – 1,433.56

5781 votes

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

DeepSeek-V4-Flash Preview (High)

1,418.31 rating

Confidence interval: 1,401.93 – 1,434.69

1476 votes

Category
spanish
Evaluated configuration
DeepSeek-V4-Flash Preview (High)
Evaluation type
human preference

Run it locally

Inference · memory estimate

DeepSeek-V4-Flash Preview (High)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license
MIT