LiveBench Story Generation: leaderboard
One of four instruction-following subtasks in LiveBench. The model writes a story about a recent Guardian article while obeying a randomly drawn subset of automatically verifiable constraints such as word limits, required keywords and formatting rules. Grading checks only whether those constraints were met, combining whole-prompt and per-instruction accuracy; the score runs from 0 to 100, higher is better.
Metric: Score. Source: livebench.ai. Saturation forecast: Not forecast. 71 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.8 Flash (High) | 82.75 | #21 (Gemini 3.8 Flash) |
| 2 | Gemini 3.7 Flash (High) | 82.25 | #23 (Gemini 3.7 Flash) |
| 3 | Qwen3.8-Flash-Next | 81.37 | #72 |
| 4 | Claude Opus 4.8 (Max) | 80.6 | #37 (Claude Opus 4.8) |
| 5 | GPT-5.4 (xHigh) | 79.02 | #79 (GPT-5.4) |
| 6 | Grok 4.6 | 78.27 | #26 |
| 7 | Claude Fable 5.1 (Max) | 78.1 | #9 (Claude Fable 5.1) |
| 8 | Muse Spark 1.3 (xHigh) | 77.48 | #24 (Muse Spark 1.3) |
| 9 | Gemini 3.1 Pro (Preview) (High) | 77.25 | #58 (Gemini 3.1 Pro (Preview)) |
| 10 | Claude Sonnet 5.5 (xHigh) | 77.08 | #11 (Claude Sonnet 5.5) |
| 11 | Qwen 3.8 Max | 76.52 | #35 |
| 12 | Qwen 3.7 Max | 76.28 | #75 |
| 13 | GPT-5.6 Sol (Max) | 75.92 | #19 (GPT-5.6 Sol) |
| 14 | Gemini 3.6 Flash (High) | 75.6 | #51 (Gemini 3.6 Flash) |
| 15 | Kimi K3 | 75.37 | #30 |
No result here: #8 Gemini 4 Argon, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=livebench-story-generation · How It Works · Data refreshed daily, snapshot 2026-10-20.