creative
Poetry and lyrics
Generate poems and lyrics with style control and variation.
#1 Recommendation
qwen-2.5-72b-instruct
Strong on Creative Writing Official (EQ-Bench Slice) creative_writing_score (78%) and Judgemark Official (EQ-Bench Slice) judgemark_score (56%)
external/qwen/qwen-2-5-72b-instruct
29.9%
Score
48.2%
Confidence
Limited benchmark evidence for this use case.
71 ranked models with average evidence of 10.9 points. Rankings may shift as more benchmark data is ingested.
Ranked Models
30
Evidence Quality
80%
Scoring
Benchmark-backed
Top Signal
Creative Writing Official (EQ-Bench Slice): creative_writing_score
All Ranked Models
| Rank | Model | Score |
|---|---|---|
| #4 | qwen-2.5-72b-instruct | 29.9% |
| #5 | gpt-4o | 28.5% |
| #9 | gemini-2.5-pro | 25.4% |
| #10 | Grok-4-0709 | 23.4% |
| #12 | gpt-4.1-20250414 | 22.8% |
| #16 | gemma-2-27b-it | 18.6% |
| #18 | xai-org/grok-4-fast-reasoning | 16.6% |
| #23 | xai-org/grok-4-1-fast-reasoning | 15.7% |
| #24 | gemini-3-pro-preview | 15.6% |
| #28 | gemini-3-flash-preview | 14.7% |
| #30 | x-ai/grok-3 | 14.5% |
| #32 | google/gemini-3.1-pro-preview | 14.2% |
| #33 | claude-sonnet-4-20250514 | 14.2% |
| #35 | gemini-2.5-flash | 14.0% |
| #40 | gpt-5-2025-08-07 | 13.1% |
| #41 | openai/gpt-5.4-2026-03-05 | 12.9% |
| #43 | gpt-5.1-2025-11-13 | 12.5% |
| #45 | anthropic/claude-sonnet-4.6 | 12.4% |
| #46 | claude-opus-4-5-20251101 | 12.3% |
| #48 | gpt-5-mini-2025-08-07 | 12.0% |
| #49 | xai-org/grok-4-1-fast-non-reasoning | 12.0% |
| #50 | Kimi-K2-Instruct | 11.9% |
| #52 | anthropic/claude-opus-4-6-thinking | 11.7% |
| #53 | gpt-5.2-2025-12-11 | 11.6% |
| #54 | gpt-4o-2024-05-13 | 11.5% |
| #57 | anthropic/claude-opus-4-5-20251101-thinking | 11.4% |
| #58 | xai-org/grok-4-fast-non-reasoning | 11.3% |
| #59 | qwen/qwen3-max | 10.9% |
| #62 | DeepSeek-V2.5 | 10.8% |
| #63 | Meta-Llama-3-70B-Instruct | 10.7% |
Compare Models
Model A leads by +1.4%
Shareable Link →Model A
qwen-2.5-72b-instruct
external/qwen/qwen-2-5-72b-instruct
Rank #4
Creative Writing Official (EQ-Bench Slice): creative_writing_score
Value 78.4% · Conf 100.0% · Weight 7.8%
artificialanalysis_creative_writing_official.creative_writing_score (Mar 12, 2026)
Judgemark Official (EQ-Bench Slice): judgemark_score
Value 55.6% · Conf 100.0% · Weight 4.2%
artificialanalysis_judgemark_official.judgemark_score (Mar 12, 2026)
EQ-Bench Leaderboard: judgemark_score
Value 55.6% · Conf 100.0% · Weight 1.9%
eq_bench.judgemark_score (Mar 12, 2026)
Galileo Agent Leaderboard v2: Avg AC
Value 76.1% · Conf 100.0% · Weight 1.9%
galileo_agent_v2.avg_ac (Mar 12, 2026)
Model B
gpt-4o
external/openai/gpt-4o
Rank #5
Creative Writing Official (EQ-Bench Slice): creative_writing_score
Value 84.4% · Conf 100.0% · Weight 8.4%
artificialanalysis_creative_writing_official.creative_writing_score (Mar 12, 2026)
Judgemark Official (EQ-Bench Slice): judgemark_score
Value 74.3% · Conf 100.0% · Weight 5.6%
artificialanalysis_judgemark_official.judgemark_score (Mar 12, 2026)
EQ-Bench Leaderboard: judgemark_score
Value 74.3% · Conf 100.0% · Weight 2.6%
eq_bench.judgemark_score (Mar 12, 2026)
MEGA-Bench: overall_score
Value 92.8% · Conf 100.0% · Weight 0.9%
mega_bench.overall_score (Mar 12, 2026)
▶Ranking Diagnostics & Missing Models
Source Lift
Ranked
71
Sources
8
Quality
Insufficient
Vals Legal Bench
vals_legal_bench
42 rows
0.7% avg lift
Vals CorpFin v2
vals_corp_fin_v2
42 rows
0.5% avg lift
Vals Tax Eval v2
vals_tax_eval_v2
41 rows
0.6% avg lift
Vals GPQA
vals_gpqa
41 rows
0.5% avg lift
Missing Strong Models
No obvious gaps right now.
▶Taxonomy Details
Core Tasks
Required Modes
Domains
Related Use Cases
creative
Screenplay scene writing
Write screenplay scenes with formatting, pacing, and strong dialogue.
Top: qwen-2.5-72b-instruct
creative
Interactive fiction / DM
Run interactive fiction with state tracking and user agency.
Top: qwen-2.5-72b-instruct
creative
NPC dialogue
Low-latency in-character dialogue suitable for games.
Top: qwen-2.5-72b-instruct
creative
SFW roleplay and simulation
Roleplay/simulations for learning or entertainment with state tracking.
Top: gemini-2.5-pro