Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-V4-Flash: 1,503.6
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-V4-Flash: 1,559.4
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4-flash: 67.27
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-flash: 73.08
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

LiveBench · Story constraint following (2025-05-30)
deepseek-v3.1-terminus: 82.33
deepseek-v3.1-terminus-thinking: 78.42
70 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-V3-0324: 1,123.9
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-V3-0324: 1,472.2
145 Models · 2026-10-25

WritingBench · Literature & arts
deepseek-v3-250324: 69.78
54 Models · 2026-10-25

WritingBench · Character design
deepseek-v3-250324: 77.2
54 Models · 2026-10-25

WritingBench · Plot development
deepseek-v3-250324: 64.73
54 Models · 2026-10-25

WritingBench · Fan fiction
deepseek-v3-250324: 70.75
54 Models · 2026-10-25

Scores from deepseek-v4-flash

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-V4-Flash: 1,503.6
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-V4-Flash: 1,559.4
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4-flash: 67.27
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-flash: 73.08
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.2: 1,514.9
145 Models · 2026-10-25

LiveBench · Story constraint following (2025-12-23)
deepseek-v3.2: 20.68
deepseek-v3.2-thinking: 55.12
65 Models · 2026-10-25

LiveBench · Theory of mind (2025-12-23)
deepseek-v3.2: 50
deepseek-v3.2-thinking: 76.92
65 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

LiveBench · Story constraint following
deepseek-v4-pro-0813: 74.93
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-pro-0813: 84.62
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-V4-Pro: 1,555.1
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-V4-Pro: 1,552.6
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4-pro: 67.75
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-pro: 80.77
71 Models · 2026-10-25

No matching public benchmark is available for this model.

Scores from deepseek-v4-pro

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-V4-Pro: 1,555.1
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-V4-Pro: 1,552.6
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4-pro: 67.75
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-pro: 80.77
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V4-Flash-0731: 1,441.3
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4-flash-0731: 70.5
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-flash-0731: 86.54
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.1: 1,436.1
145 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

EQ-Bench 3
deepseek-ai/DeepSeek-R1: 1,292.4
83 Models · 2026-10-25

Creative Writing v3
deepseek-ai/DeepSeek-R1: 1,500
145 Models · 2026-10-25

WritingBench · Literature & arts
deepseek-r1: 71.93
54 Models · 2026-10-25

WritingBench · Character design
deepseek-r1: 72
54 Models · 2026-10-25

WritingBench · Plot development
deepseek-r1: 67.27
54 Models · 2026-10-25

WritingBench · Fan fiction
deepseek-r1: 69
54 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

LiveBench · Story constraint following
deepseek-v4-flash-vision-exp: 73.83
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4-flash-vision-exp: 78.85
71 Models · 2026-10-25

Scores from deepseek-v3.1

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.1: 1,436.1
145 Models · 2026-10-25

Scores from deepseek-v3.2

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.2: 1,514.9
145 Models · 2026-10-25

LiveBench · Story constraint following (2025-12-23)
deepseek-v3.2: 20.68
deepseek-v3.2-thinking: 55.12
65 Models · 2026-10-25

LiveBench · Theory of mind (2025-12-23)
deepseek-v3.2: 50
deepseek-v3.2-thinking: 76.92
65 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-v4.1-flash: 1,539.9
145 Models · 2026-10-25

LiveBench · Story constraint following
deepseek-v4.1-flash-max: 72.68
71 Models · 2026-10-25

LiveBench · Theory of mind
deepseek-v4.1-flash-max: 80.77
71 Models · 2026-10-25

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-R1-0528: 1,421.5
145 Models · 2026-10-25

LiveBench · Story constraint following (2025-05-30)
deepseek-r1-0528: 81.25
70 Models · 2026-10-25

WritingBench · Literature & arts
deepseek-r1-0528: 80.91
54 Models · 2026-10-25

WritingBench · Character design
deepseek-r1-0528: 86
54 Models · 2026-10-25

WritingBench · Plot development
deepseek-r1-0528: 82
54 Models · 2026-10-25

WritingBench · Fan fiction
deepseek-r1-0528: 84
54 Models · 2026-10-25