Scale AI - SEAL Platform: leaderboard
Evaluation platform hosting 15+ benchmarks across agentic, frontier, safety, and multimodal categories. The enterprise standard for model evaluation.
Metric: Score. Source: scale.com. Status: years away from saturation. 28 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 Mini | 1237 |
| 2 | O3 Mini | 1137 |
| 3 | GPT-4o | 1132 |
| 4 | O1 Preview | 1123 |
| 5 | Gemini 2.0 Flash | 1111 |
| 6 | Gemini 2.0 Flash (Thinking) | 1108 |
| 7 | DeepSeek R1 | 1100 |
| 8 | O1 (2024-12-17) | 1083 |
| 9 | Claude 3.5 Sonnet (20240620) | 1079 |
| 10 | GPT-4o (2024-08-06) | 1045 |
| 11 | GPT-4o (2024-05-13) | 1036 |
| 12 | GPT-4 Turbo (Preview) | 1034 |
| 13 | Mistral Large 2 (Jul) | 1029 |
| 14 | Llama 3.1 405B Instruct | 1022 |
| 15 | Gemini 1.5 Pro (0514) | 994 |
Interactive version: theaggregate.ai/benchmark?slug=scale-ai-seal-platform · How It Works · Data refreshed daily, snapshot 2026-09-05.