SciKnowEval: leaderboard
Scientific knowledge evaluation benchmark spanning biology, chemistry, materials science, and physics tasks.
Metric: Inverse Overall Rank. Source: github.com. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 26 | #299 |
| 2 | GPT-4o (2024-05-13) | 25 | #325 |
| 3 | GPT-4 Turbo | 23 | #404 |
| 4 | Llama 3 70B Instruct | 21 | #634 |
| 5 | Qwen Max | 20 | #382 |
| 6 | Claude 3 Sonnet (20240229) | 19 | #677 |
| 7 | Qwen 1.5 14B Chat | 16 | #977 |
| 8 | GPT-3.5 Turbo (0125) | 15 | #984 |
| 9 | Llama 3 8B Instruct | 14 | #1129 |
| 10 | Qwen 1.5 7B Chat | 10 | #1324 |
| 11 | chatglm3-6B | 7 | #1439 |
| 12 | Llama 2 13B Chat | 5 | #1465 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sciknoweval · How It Works · Data refreshed daily, snapshot 2026-10-20.