DeepSeek R1 is cheaper than GPT-4.1 at $0.55/M vs $2/M input tokens.
| Spec | DeepSeek R1 | GPT-4.1 |
|---|---|---|
| Input price | $0.55/M tokens | $2/M tokens |
| Output price | $2.19/M tokens | $8/M tokens |
| Context window | 128K tokens | 1.0M tokens |
| Parameters | 671B (37B active) | Not disclosed |
| Weights | Open | Closed |
| Free API (OpenRouter) | No | No |
| Released | Feb 2025 | Apr 2025 |
| SWE-bench Verified | 49.2% | 54.6% |
Not enough votes to call it. On the specs, GPT-4.1 has the edge: bigger model tier, newer, bigger context window.
DeepSeek R1 takes Reasoning. DeepSeek R1 costs 3.7x less per token.
DeepSeek R1 is cheaper on both: 3.6× input, 3.7× output.
Per million tokens. Prices and uptime via OpenRouter, checked 8 Sep 2026.
DeepSeek R1 uses 2.5x more transitions