DeepSeek-V2.5
Stored tensors can include buffers and repeated shared weights. A verified unique parameter count is unavailable. Checkpoint size is not the memory required to run the model.
GeekRadar / The comparison desk
COMPARE.
AI modelsDeepSeek-V2.5
AI model
- Context window
- 163,840 tokens
- Open weights
- Yes
Capability & preference
Human preference · Arena
Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.
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-V2.5
1,281.23 rating
Confidence interval: 1,268.19 – 1,294.26
2242 votes
- Category
- chinese
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,301.04 rating
Confidence interval: 1,291.75 – 1,310.33
4252 votes
- Category
- coding
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,236.47 rating
Confidence interval: 1,225.99 – 1,246.95
3450 votes
- Category
- creative_writing
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,284.23 rating
Confidence interval: 1,278.08 – 1,290.39
12845 votes
- Category
- english
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,192.56 rating
Confidence interval: 1,185.09 – 1,200.03
15337 votes
- Category
- exclude_ties
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / experthuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,239.53 rating
Confidence interval: 1,224.79 – 1,254.27
1523 votes
- Category
- expert
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,289.47 rating
Confidence interval: 1,251.12 – 1,327.82
222 votes
- Category
- french
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,226.43 rating
Confidence interval: 1,201.91 – 1,250.96
563 votes
- Category
- german
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,270.18 rating
Confidence interval: 1,262.66 – 1,277.70
6869 votes
- Category
- hard_prompts
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,279.61 rating
Confidence interval: 1,270.34 – 1,288.88
4055 votes
- Category
- hard_prompts_english
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,248.13 rating
Confidence interval: 1,236.67 – 1,259.60
2825 votes
- Category
- industry_business_and_management_and_financial_operations
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,229.89 rating
Confidence interval: 1,219.89 – 1,239.88
3664 votes
- Category
- industry_entertainment_and_sports_and_media
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,282.21 rating
Confidence interval: 1,266.78 – 1,297.64
1518 votes
- Category
- industry_legal_and_government
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,273.14 rating
Confidence interval: 1,263.61 – 1,282.67
4318 votes
- Category
- industry_life_and_physical_and_social_science
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,273.46 rating
Confidence interval: 1,262.99 – 1,283.92
3039 votes
- Category
- industry_mathematical
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,259.15 rating
Confidence interval: 1,241.83 – 1,276.48
1214 votes
- Category
- industry_medicine_and_healthcare
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,294.98 rating
Confidence interval: 1,287.02 – 1,302.94
6692 votes
- Category
- industry_software_and_it_services
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,246.78 rating
Confidence interval: 1,238.59 – 1,254.97
6613 votes
- Category
- industry_writing_and_literature_and_language
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,249.00 rating
Confidence interval: 1,242.32 – 1,255.67
10175 votes
- Category
- instruction_following
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,190.18 rating
Confidence interval: 1,163.44 – 1,216.92
479 votes
- Category
- japanese
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,209.42 rating
Confidence interval: 1,176.09 – 1,242.76
344 votes
- Category
- korean
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,272.73 rating
Confidence interval: 1,262.65 – 1,282.81
3499 votes
- Category
- longer_query
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,270.67 rating
Confidence interval: 1,260.88 – 1,280.46
3649 votes
- Category
- math
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,260.66 rating
Confidence interval: 1,251.03 – 1,270.29
4473 votes
- Category
- multi_turn
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,253.12 rating
Confidence interval: 1,246.80 – 1,259.45
11727 votes
- Category
- non_english
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,271.23 rating
Confidence interval: 1,266.55 – 1,275.91
24572 votes
- Category
- overall
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,252.60 rating
Confidence interval: 1,242.33 – 1,262.87
3519 votes
- Category
- russian
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating
DeepSeek-V2.5
1,248.14 rating
Confidence interval: 1,213.93 – 1,282.35
293 votes
- Category
- spanish
- Evaluated configuration
- DeepSeek-V2.5
- Evaluation type
- human preference
Arena text preference / chinesehuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,281.28 rating
Confidence interval: 1,268.24 – 1,294.31
2242 votes
- Category
- chinese
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,301.14 rating
Confidence interval: 1,291.85 – 1,310.44
4252 votes
- Category
- coding
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,236.62 rating
Confidence interval: 1,226.14 – 1,247.10
3450 votes
- Category
- creative_writing
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,284.18 rating
Confidence interval: 1,278.03 – 1,290.33
12845 votes
- Category
- english
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,192.57 rating
Confidence interval: 1,185.10 – 1,200.04
15337 votes
- Category
- exclude_ties
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / experthuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,239.57 rating
Confidence interval: 1,224.83 – 1,254.32
1523 votes
- Category
- expert
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,289.16 rating
Confidence interval: 1,250.79 – 1,327.52
222 votes
- Category
- french
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,226.43 rating
Confidence interval: 1,201.95 – 1,250.92
563 votes
- Category
- german
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,270.17 rating
Confidence interval: 1,262.65 – 1,277.69
6869 votes
- Category
- hard_prompts
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,279.49 rating
Confidence interval: 1,270.22 – 1,288.77
4055 votes
- Category
- hard_prompts_english
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,248.17 rating
Confidence interval: 1,236.71 – 1,259.64
2825 votes
- Category
- industry_business_and_management_and_financial_operations
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,229.94 rating
Confidence interval: 1,219.95 – 1,239.94
3664 votes
- Category
- industry_entertainment_and_sports_and_media
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,281.95 rating
Confidence interval: 1,266.52 – 1,297.39
1518 votes
- Category
- industry_legal_and_government
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,273.12 rating
Confidence interval: 1,263.59 – 1,282.65
4318 votes
- Category
- industry_life_and_physical_and_social_science
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,273.45 rating
Confidence interval: 1,262.99 – 1,283.91
3039 votes
- Category
- industry_mathematical
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,259.21 rating
Confidence interval: 1,241.88 – 1,276.54
1214 votes
- Category
- industry_medicine_and_healthcare
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,295.06 rating
Confidence interval: 1,287.10 – 1,303.02
6692 votes
- Category
- industry_software_and_it_services
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,246.84 rating
Confidence interval: 1,238.65 – 1,255.03
6613 votes
- Category
- industry_writing_and_literature_and_language
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,249.01 rating
Confidence interval: 1,242.33 – 1,255.68
10175 votes
- Category
- instruction_following
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,190.01 rating
Confidence interval: 1,163.32 – 1,216.70
479 votes
- Category
- japanese
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,209.46 rating
Confidence interval: 1,176.14 – 1,242.79
344 votes
- Category
- korean
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,272.77 rating
Confidence interval: 1,262.69 – 1,282.85
3499 votes
- Category
- longer_query
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,270.63 rating
Confidence interval: 1,260.84 – 1,280.42
3649 votes
- Category
- math
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,260.65 rating
Confidence interval: 1,251.02 – 1,270.28
4473 votes
- Category
- multi_turn
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,253.22 rating
Confidence interval: 1,246.89 – 1,259.54
11727 votes
- Category
- non_english
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,271.21 rating
Confidence interval: 1,266.53 – 1,275.89
24572 votes
- Category
- overall
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,252.76 rating
Confidence interval: 1,242.49 – 1,263.03
3519 votes
- Category
- russian
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating
DeepSeek-V2.5
1,248.06 rating
Confidence interval: 1,213.83 – 1,282.29
293 votes
- Category
- spanish
- Evaluated configuration
- deepseek-v2.5
- Evaluation type
- human preference
Run it locally
Capabilities
- Context window
- 163,840 tokens
- Open weights
- Yes
- Model license
- other
- Checkpoint tensor payload
- 439.10 GiB
Input modalities
- text
Output modalities
- text
Processing & platform
- Architecture
- deepseek_v2
0 selected Compare 0