BenchLeader

EQ-Bench Creative Writing v3 leaderboard

As of 16 Nov 2026, the top score on EQ-Bench Creative Writing v3 is 2177 by GPT-6 Astra (OpenAI), ahead of Kimi K3.1 Pro at 2173, across 125 model configurations with a published result (EQ-Bench). Short creative-writing pieces from 32 prompts, compared pairwise by an LLM judge and turned into an Elo-style rating. EQ-Bench Creative Writing v3 is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 16 Nov 2026)

  • GPT-6 Astra holds the top EQ-Bench Creative Writing v3 score at 2177 (EQ-Bench, 16 Nov 2026).
  • Kimi K3.1 Pro is second on EQ-Bench Creative Writing v3 at 2173 (EQ-Bench, 16 Nov 2026).
  • EQ-Bench Creative Writing v3: Short creative-writing pieces from 32 prompts, compared pairwise by an LLM judge and turned into an Elo-style rating (EQ-Bench).
  • 125 model configurations have a published EQ-Bench Creative Writing v3 result on BenchLeader (BenchLeader, 16 Nov 2026).
  • EQ-Bench Creative Writing v3 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Short creative-writing pieces from 32 prompts, compared pairwise by an LLM judge and turned into an Elo-style rating.

Published by
EQ-Bench
Category
Human preference
Index weight
Reference only
Models
125
Data as of
16 Nov 2026

MIT — EQ-Bench.

125 of 125
#
1GPT-6 Astra217766.4
2Kimi K3.1 Pro217361.8
3Claude Fable 5.1215566.4
4Claude Opus 5213563.2
5GPT-6 Sol2126–
6Kimi K3208360.7
7GLM 5.32076–
8Claude Opus 5.5205161.0
9Grok 4.72012–
10Muse Wonder200764.4
11GPT-5.6 Sol197454.3
12Ox Alpha1974–
13Claude Fable 5194465.4
14Muse Spark 1.1192862.2
15Claude Opus 4.7191563.1
16Muse Spark 1.3190761.6
17GPT-5.6 Terra1856–
18GPT-5.5184561.0
19Qwen3.8 2.4T A95B184451.4
20Muse Spark 1.21842–
21GPT-5.4184157.9
22Claude Opus 4.8184058.7
23GPT-5.6 Luna183247.2
24Claude Sonnet 4.6181157.6
25Claude Opus 4.6181062.1
26Muse Glimmer180454.4
27Claude Sonnet 5179853.0
28GLM 5.2175951.9
29Gemini 4 Argon175864.4
30Gemini 3.8 Flash174456.4
31Kimi K2.6172155.1
32Gemini 3.7 Flash1720–
33GPT-5.2170356.7
34Nemotron 3 Ultra 550B A55B NVFP41692–
35GPT-5.3 Chat169058.3
36Claude Opus 4.5168657.1
37Claude Sonnet 4.5167753.8
38o3167560.4
39Qwen3.8 27B167155.1
40Kimi K2166651.1
41GPT-5.4 Mini1665–
42Kimi K2 Thinkingthinking163152.3
43GPT-5162656.6
44Inkling160950.4
45Gemini 3.6 Flash160249.3
46GLM-5160154.9
47GLM 5.1159455.4
48Kimi K2.5157954.6
49Claude Opus 4157851.9
50Grok 4.5157659.3
51Grok 4.201574–
52DeepSeek V4 Flash155953.8
53Gemini 3.5 Flash Lite155951.4
54DeepSeek V4 Pro155353.0
55DeepSeek V4.1 Flash1540–
56Gemini 3 Pro152460.0
57DeepSeek V3.2151550.2
58GPT-4o150142.8
59DeepSeek R1150047.2
60MiMo-V2.5-Pro149355.5

More human preference benchmarks · All LLM benchmarks

Cite as: BenchLeader, “EQ-Bench Creative Writing v3 leaderboard”, data as of 16 Nov 2026, https://www.benchleader.com/benchmarks/eqbench_creative_writing.

EQ-Bench Creative Writing v3: questions

What does EQ-Bench Creative Writing v3 measure?
Short creative-writing pieces from 32 prompts, compared pairwise by an LLM judge and turned into an Elo-style rating. Scores are reported in Elo-style ratings from pairwise comparisons; higher is better.
What is the top score on EQ-Bench Creative Writing v3?
As of 16 Nov 2026, the highest score on EQ-Bench Creative Writing v3 is 2177 by GPT-6 Astra, ahead of Kimi K3.1 Pro at 2173. Scores come from EQ-Bench and are re-read daily.
Which AI model leads EQ-Bench Creative Writing v3?
GPT-6 Astra leads EQ-Bench Creative Writing v3 with 2177 as of 16 Nov 2026, ahead of Kimi K3.1 Pro at 2173.
How many models have EQ-Bench Creative Writing v3 results?
125 model configurations have a EQ-Bench Creative Writing v3 result on BenchLeader, all taken from EQ-Bench.
Who runs EQ-Bench Creative Writing v3 and how often is it updated?
EQ-Bench Creative Writing v3 is published by EQ-Bench. BenchLeader re-reads the published results once a day and records the date each result was published.
Does EQ-Bench Creative Writing v3 count toward the BenchLeader Index?
No. EQ-Bench Creative Writing v3 is shown for reference but left out of the composite index.