Poetry and lyrics
Generate poems and lyrics with style control and variation.
Provisional leader
Grok-4-0709
Current leader based on limited benchmark evidence. Treat this ranking as directional until coverage improves.
31.7%
Best benchmark score
41.4%
Confidence
All ranked models — top 3
Ranked Models
30
Evidence Quality
82%
Evidence Points
23
Top Signal
UGI Leaderboard: Writing ✍️
All Ranked Models
| Rank | Model | Score |
|---|---|---|
| 🥇 | Grok-4-0709 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 31.7% |
| 🥈 | gemini-2.5-pro Strong on UGI Leaderboard Writing ✍️ and MWS Vision Bench validation_overall_score | 31.3% |
| 🥉 | gpt-5-2025-08-07 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 28.6% |
| #4 | gemini-3-pro-preview Strong on UGI Leaderboard Writing ✍️ and BFCL Memory Official Memory Acc | 28.3% |
| #5 | claude-sonnet-4 Strong on UGI Leaderboard Writing ✍️ and Galileo Agent Leaderboard v2 Avg AC | 27.5% |
| #6 | o3-20250416 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 27.2% |
| #7 | gemini-3.1-pro-preview Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 27.1% |
| #8 | gpt-4.1-20250414 Strong on UGI Leaderboard Writing ✍️ and Galileo Agent Leaderboard v2 Avg AC | 26.7% |
| #9 | gemini-3-flash-preview Strong on UGI Leaderboard Writing ✍️ and MWS Vision Bench validation_overall_score | 26.1% |
| #11 | gpt-5.2-2025-12-11 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 23.6% |
| #12 | gpt-5.4-2026-03-05 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 23.1% |
| #13 | claude-sonnet-4.6 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 22.7% |
| #14 | grok-4-1-fast-reasoning Strong on UGI Leaderboard Writing ✍️ and BFCL Memory Official Memory Acc | 22.5% |
| #15 | qwen-2.5-72b-instruct Strong on EQ-Bench Leaderboard judgemark_score and Galileo Agent Leaderboard v2 Avg AC | 22.0% |
| #17 | claude-opus-4-5-20251101 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 21.5% |
| #18 | kimi-k2.5-thinking Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 21.3% |
| #19 | gpt-5.1-2025-11-13 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 20.2% |
| #20 | gpt-5-mini-2025-08-07 Strong on MWS Vision Bench validation_overall_score and Vals MedQA overall_accuracy_pct | 19.8% |
| #21 | o4-mini Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 19.8% |
| #23 | gpt-4o Strong on EQ-Bench Leaderboard judgemark_score and MEGA-Bench overall_score | 19.7% |
| #24 | gemini-2.5-flash Strong on MWS Vision Bench validation_overall_score and Galileo Agent Leaderboard v2 Avg TSQ | 19.2% |
| #25 | grok-4-fast-reasoning Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 18.9% |
| #27 | Kimi-K2-Instruct Strong on UGI Leaderboard Entertainment and UGI Leaderboard Writing ✍️ | 18.0% |
| #28 | Kimi K2 Thinking Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 18.0% |
| #30 | grok-3 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 16.8% |
| #33 | grok-4-1-fast-non-reasoning Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 16.3% |
| #34 | claude-opus-4 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 16.1% |
| #35 | claude-opus-4-1-20250805 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 16.1% |
| #37 | GLM-4.6 Strong on UGI Leaderboard Entertainment and UGI Leaderboard Writing ✍️ | 15.6% |
| #38 | claude-opus-4-6 Strong on UGI Leaderboard Writing ✍️ and UGI Leaderboard Entertainment | 15.4% |
Compare Models
▶Ranking diagnostics & missing models
Source lift
Ranked
82
Sources
8
Quality
Low
UGI Leaderboard
Vals Legal Bench
Vals GPQA
Vals Tax Eval v2
Missing frontier models
No obvious gaps right now.
▶Taxonomy & task details
Core tasks
Required modes
Domains
Related in Creative
Interactive fiction / DM
Run interactive fiction with state tracking and user agency.
NPC dialogue
Low-latency in-character dialogue suitable for games.
SFW roleplay and simulation
Roleplay/simulations for learning or entertainment with state tracking.
Screenplay scene writing
Write screenplay scenes with formatting, pacing, and strong dialogue.