LisanBench: leaderboard
LisanBench evaluates LLMs on open-ended word-chain generation, stressing vocabulary, planning, recall, persistence, and exact edit-distance rule following.
Metric: Mean Path Length / Current Maximum. Source: github.com. Source: lisanbench.com. Saturation forecast: Around January 2027. 154 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 (High) | 1 | #12 (Claude Opus 5) |
| 2 | Claude Opus 4.7 (xHigh) | 0.93 | #49 (Claude Opus 4.7) |
| 3 | Claude Fable 5 (Medium) | 0.89 | #16 (Claude Fable 5) |
| 4 | Kimi K3 (Max) | 0.68 | #30 (Kimi K3) |
| 5 | GPT-5.5 (Medium) | 0.64 | #31 (GPT-5.5) |
| 6 | Claude Opus 4.8 (High) | 0.61 | #37 (Claude Opus 4.8) |
| 7 | Claude Opus 5 (Medium) | 0.53 | #12 (Claude Opus 5) |
| 8 | Gemini 3.1 Pro (Preview) (High) | 0.42 | #58 (Gemini 3.1 Pro (Preview)) |
| 9 | Claude Sonnet 5 (High) | 0.37 | #57 (Claude Sonnet 5) |
| 10 | GPT-5.4 (Medium) | 0.35 | #79 (GPT-5.4) |
| 11 | GPT-5.6 Sol (Medium) | 0.3 | #19 (GPT-5.6 Sol) |
| 12 | Gemini 3 Pro (Preview) (High) | 0.3 | #66 (Gemini 3 Pro (Preview)) |
| 13 | Gemini 3.5 Flash (High) | 0.27 | #60 (Gemini 3.5 Flash) |
| 14 | GPT-5.2 (Medium) | 0.23 | #110 (GPT-5.2) |
| 15 | Gemini 3.1 Pro (Preview) (Low) | 0.23 | #58 (Gemini 3.1 Pro (Preview)) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lisanbench · How It Works · Data refreshed daily, snapshot 2026-10-20.