LiveBench Plot Unscrambling — leaderboard
Metric: Score. Source: livebench.ai. 73 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 55.14 |
| 2 | Claude 3 Opus (20240229) | 50.16 |
| 3 | Llama 3.1 405B Instruct | 47.38 |
| 4 | GPT-4 (0613) | 45.54 |
| 5 | GPT-4o (2024-08-06) | 45.11 |
| 6 | GPT-4 Preview (0125) | 43.49 |
| 7 | Gemini 1.5 Pro (Preview 0827) | 43.44 |
| 8 | GPT-4o (2024-05-13) | 42.81 |
| 9 | GPT-4o ChatGPT | 42.53 |
| 10 | Gemini 1.5 Pro (Preview 0801) | 41.87 |
| 11 | Mistral Large 2 (Jul) | 40.05 |
| 12 | Gemini 1.5 Pro (0514) | 39.92 |
| 13 | GPT-4 Turbo | 39.62 |
| 14 | Llama 3.1 70B Instruct | 39.42 |
| 15 | Hermes 3 - Llama-3.1 70B | 37.32 |
Interactive version: theaggregate.ai/benchmark?slug=livebench-plot-unscrambling · How the rankings work · Data refreshed daily, snapshot 2026-07-22.