Skip to content

Benchmark

Arena ELO - Creative Writing

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.14

Score distribution

1,287.4
Mean
1,318.1
Median
144.2
Std dev
1,079.9
10th percentile
1,442.3
90th percentile
793.9
Min
793.91,532.7
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.14 for chat models, measured over 364 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51DeepSeek V4 ProHigh, 20260813· LMArena data1,438
52Qwen3.7 Plus· LMArena data1,437.8
53GLM 5· LMArena data1,437.4
54GPT 5.4High· LMArena data1,437.3
55MiMo V2.5 Pro· LMArena data1,437.1
56Grok 4.20 Multi-Agent0309, beta· LMArena data1,436.7
57DeepSeek V4.1 Flash 0910Max· LMArena data1,436.7
58Kimi K3.1 ProMax· LMArena data1,436.4
59Gemini 3 FlashMinimal· LMArena data1,435.3
60Muse Spark 1.1· LMArena data1,435.2
61Qwen3.6 Max Preview· LMArena data1,434.7
62Claude Sonnet 4.6· LMArena data1,434.6
63Kimi K2.6· LMArena data1,433.5
64Grok 4.20Reasoning, 0309, beta· LMArena data1,431.9
65Grok 4.6High· LMArena data1,427.7
66GPT 5.1High· LMArena data1,426.8
67ERNIE 5.00110· LMArena data1,425.9
68GPT 5.4· LMArena data1,425.9
69Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,424.1
70Kimi K2.5Thinking· LMArena data1,421.8
71GPT-6 AstraMax· LMArena data1,421.4
72ERNIE-5.0-Preview1203· LMArena data1,421.3
73Gemini 3.5 Flash Lite· LMArena data1,421.1
74GPT-6.1 SolMax· LMArena data1,421
75GPT-5.5 Instant· LMArena data1,418.1
76ERNIE-5.0-Preview1022· LMArena data1,417.6
77GPT-6.2 AstraMax· LMArena data1,417
78Step 5 Preview· LMArena data1,416.2
79MiMo V2 Pro· LMArena data1,416.2
80Gemma 4 31B IT· LMArena data1,414.7
81Grok 3 Preview02-24· LMArena data1,414.2
82Claude Sonnet 5High· LMArena data1,413.9
83GPT-5.6 TerraxHigh· LMArena data1,413.6
84GLM 4.6· LMArena data1,412
85Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,411.8
86Grok 4.1· LMArena data1,411.2
87Claude Opus 4.120250805· LMArena data1,410.3
88Doubao Seed 2.0 Pro· LMArena data1,407.3
89DeepSeek R10528· LMArena data1,405.4
90ChatGPT-4o Latest20250326· LMArena data1,405.1
91Hy3 (Deep Infra)· LMArena data1,404.7
92Grok 4.1Thinking· LMArena data1,404.1
93Qwen3.6 Plus· LMArena data1,403.8
94DeepSeek V3.1 Terminus· LMArena data1,403.5
95MiMo V2.6 Flash· LMArena data1,403.4
96GPT 5.1· LMArena data1,403.3
97DeepSeek V3.2 Exp· LMArena data1,403
98GPT-6.2 SolMax· LMArena data1,402.9
99DeepSeek V4 Flash· LMArena data1,402.1
100Qwen3.5 397B A17B· LMArena data1,402

Showing 51–100 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)