o3 Beats Humans at Math and Science
Competition Math (AIME 2024) accuracy 100 80 60 40 20 0 o1 preview 56.7 o1 83.3 o3 96.7 PhD-level Science Questions (GPQA Diamond) accuracy 100 80 60 40 20 0 o1 preview 78.3 o1 78.0 o3 87.7
Two bar charts compare o1 preview, o1, and o3 on competition mathematics and PhD-level science questions. The joke is that o3 dramatically outperforms earlier models, humorously presenting AI benchmark results as an impressive flex for people discussing OpenAI models.
- @dev_meme / 6469 2024-12-20
- Type
- image
- Size
- 800×453 · 22.5 KiB
- Mood
- sarcastic, impressed
- Language
- en
- Added
