Story Theory Bench: leaderboard
Creative narrative generation benchmark: LLMs write short stories incorporating 10 required elements (characters, objects, concepts, attributes, motivations). Scored 50% programmatic + 50% LLM judge across 34 tasks.
Metric: Score (%). Source: github.com. Saturation forecast: Around January 2027. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 99.6 | #79 |
| 2 | GLM-5 | 99.6 | #141 |
| 3 | Mercury 2 | 99.1 | #433 |
| 4 | Qwen 3.5 27B | 98.9 | #187 |
| 5 | Qwen 3.5 Flash (02-23) | 98.6 | #185 |
| 6 | Gemini 3.1 Flash Lite (Preview) | 98.4 | #194 |
| 7 | Claude Opus 4.6 | 98.1 | #63 |
| 8 | DeepSeek V3.2 | 92.2 | #205 |
| 9 | Claude Opus 4.5 | 90.9 | #83 |
| 10 | Claude Sonnet 4.5 | 90.1 | #142 |
| 11 | Step 3.5 Flash | 90.1 | #270 |
| 12 | Claude Sonnet 4 | 89.6 | #199 |
| 13 | O3 | 89.5 | #124 |
| 14 | GLM-4.7 | 88.8 | #192 |
| 15 | Kimi K2 (Thinking) | 88.7 | #243 (Kimi K2) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=story-theory-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.