FLY-EVAL++ - Single-Step Prediction - Total: leaderboard
Metric: S1 total score (%). Source: arxiv.org. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 87.77 |
| 2 | DeepSeek R1 0528 | 87.76 |
| 3 | GPT-5 | 87.69 |
| 4 | O4 Mini | 87.41 |
| 5 | Kimi K2 (Thinking) | 86.68 |
| 6 | Gemini 2.5 Pro | 86.12 |
| 7 | Llama 3.1 405B | 83.33 |
| 8 | Kimi K2 0905 | 82.8 |
| 9 | Claude Sonnet 4.5 | 82.5 |
| 10 | Grok 4 | 82.41 |
| 11 | DeepSeek V3.1 | 82.36 |
| 12 | Claude 3.7 Sonnet (20250219) | 82.36 |
| 13 | GLM-4.6 | 82.11 |
| 14 | DeepSeek V3.2 Exp | 81.92 |
| 15 | DeepSeek V3 | 81.71 |
Interactive version: theaggregate.ai/benchmark?slug=fly-eval-plus-plus-single-step-prediction-total · How It Works · Data refreshed daily, snapshot 2026-09-19.