NARRA-Gym for Evaluating Interactive Narrative: leaderboard
Interactive narrative tasks require LLMs to sustain a coherent, evolving story while adapting to a user over multiple turns. However, suitable benchmarks for this setting are limited: existing evaluations often focus on static prompts, isolated story generations, or post-hoc ratings, and therefore miss whether models can jointly manage story generation, long-context state and pacing, character sim
Metric: StoryQ (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 3.9 | #90 |
| 2 | Claude Opus 4.6 | 3.48 | #63 |
| 3 | DeepSeek V4 Pro | 3.2 | #102 |
| 4 | GPT-5.4 | 3.17 | #79 |
| 5 | DeepSeek V3.2 | 3.14 | #205 |
| 6 | GLM-5.1 | 2.68 | #97 |
| 7 | Gemini 3.1 Pro (Preview) | 2.56 | #58 |
| 8 | Qwen 3.5 397B A17B | 2.43 | #146 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=narra-gym-for-evaluating-interactive-narrative · How It Works · Data refreshed daily, snapshot 2026-10-20.