# Scale SEAL Leaderboards

Scale AI tests AI models on private, expert-written questions that models cannot have memorised, across areas like coding, reasoning and safety. One representative board (Humanity's Last Exam, Diamond) is shown.

Category: aggregators. Type: leaderboard. Status: active.
Metric: Share of expert-level questions answered correctly in Humanity's Last Exam (Diamond) (%) (higher is better).
Source: Scale AI (SEAL) — https://labs.scale.com/leaderboard/hle-diamond, results as of 2026-09-23, retrieved 2026-11-17, license: not stated.

| Rank | Model | Company | Score (%) | Confidence interval (+/-) |
|---|---|---|---|---|
| 1 | claude-fable-5-2 | anthropic | 72.4 | 2.8 |
| 2 | gpt-6-astra | openai | 60.6 | 3 |
| 3 | claude-opus-5-5 | anthropic | 55 | 3.1 |
| 4 | gemini-4-argon | google | 53.2 | 3.07 |
| 5 | claude-fable-5-1 | anthropic | 51.3 | 3.1 |
| 6 | claude-opus-5 | anthropic | 38.6 | 3 |

Notes: SEAL hosts ~30 leaderboards; only HLE (Diamond) shown, non-deprecated entries. as_of is the board's updatedAt field; newest entries dated Nov 2026. Some entries carry a contamination warning on the site. Data read from page-embedded JSON.
