Best Benchmark Score
o3
48.5/100
Model Specifications
| Spec | DeepSeek-R1 | o3 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Release Date | Jan 2025 | Apr 2025 |
| Knowledge Cutoff | — | 2024-05-31 |
| Parameters | 671000000000B | Undisclosed |
| Context Window | — | — |
| Max Output | — | — |
| Open Source | Yes | No |
| License | mit | proprietary |
| Tokenizer | — | — |
| Modality | — | — |
| Reasoning Model |
Pricing Comparison
| Price (per 1M tokens) | DeepSeek-R1 | o3 |
|---|
API Provider Pricing
Capabilities
| Feature | DeepSeek-R1 | o3 |
|---|---|---|
| Text Input | ||
| Image Input (Vision) | ||
| Audio Input | ||
| Video Input |
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | DeepSeek-R1 | o3 |
|---|---|---|
| SWE-bench Verified | — | 69.1 |
math
| Benchmark | DeepSeek-R1 | o3 |
|---|---|---|
| AIME 2025 | — | 86.4 |
| FrontierMath | — | 15.8 |
reasoning
Category Winners
coding
59.1
math
43.4
reasoning
35.1
multimodal
74.7
agent
23.1