DeepSeek
DeepSeek-R1
Open-weight reasoning model under the MIT licence. DeepSeek said it performed on par with OpenAI's o1 (self-reported).
- Announced
- Access
- Public
Benchmark scores
| Test | Score | Best on the tracker | How it was measured |
|---|---|---|---|
| GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated | 72% | Gemini 4 Argon 96% | Independentdefault settings (DeepSeek-R1)Run 26 May 2025 |
| ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated | 1.3% | GPT-6 Astra 95% | IndependentARC Prize config "Deepseek R1" (R1) |
| OTIS Mock AIMEcompetition mathsSaturated | 53% | GPT-5.5 100% | Independentdefault settings (DeepSeek-R1)Run 26 Feb 2025 |
Sources
- DeepSeek-R1 ReleaseDeepSeek · Official · api-docs.deepseek.com
- Epoch AI Benchmarking Hub (model release dates and runs)Epoch AI · Dataset · 9 Nov 2026 · epoch.ai