NYT Connections Original: leaderboard
Original version: 436 standard NYT Connections puzzles without modifications. Models get one attempt (humans get four). ~71% average human solve rate.
Metric: Score (%). Source: github.com. Saturation forecast: Estimated already saturated. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 | 90.7 | #183 |
| 2 | O1 Preview | 87.1 | #193 |
| 3 | O3 Mini | 72.4 | #272 |
| 4 | DeepSeek R1 | 54.4 | #251 |
| 5 | O1 Mini | 42.2 | #323 |
| 6 | Gemini 2.0 Flash (01-21) (Thinking) | 37 | #322 (Gemini 2.0 Flash (01-21)) |
| 7 | GPT-4 Turbo | 28.3 | #404 |
| 8 | GPT-4o (2024-11-20) | 27.9 | #381 |
| 9 | GPT-4o (2024-08-06) | 26.5 | #336 |
| 10 | Llama 3.1 405B | 26.3 | #466 |
| 11 | Claude 3.5 Sonnet (20241022) | 25.9 | #291 |
| 12 | Claude 3 Opus | 24.8 | #571 |
| 13 | Llama 3.3 70B | 23.7 | #579 |
| 14 | Grok Beta | 23.7 | #487 |
| 15 | Gemini 1.5 Pro (Sept) | 22.7 | #448 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=nyt-connections-original · How It Works · Data refreshed daily, snapshot 2026-10-20.