DeepSeek V3.2 is cheaper than GPT-4.5 at $0.28/M vs $75/M input tokens.
| Spec | DeepSeek V3.2 | GPT-4.5 |
|---|---|---|
| Input price | $0.28/M tokens | $75/M tokens |
| Output price | $0.42/M tokens | $150/M tokens |
| Context window | 131K tokens | 128K tokens |
| Weights | Open | — |
| Free API (OpenRouter) | No | No |
| Released | Dec 2025 | Feb 2025 |
| SWE-bench Verified | 73.1% | 28.0% |
Not enough votes to call it. On the specs, nothing separates them.
DeepSeek V3.2 costs 357x less per token.
DeepSeek V3.2 is cheaper on both: 268× input, 357× output.
No hosts listed on OpenRouter.
Per million tokens. Prices and uptime via OpenRouter, checked 9 Sep 2026.
GPT-4.5 uses 4.3x more transitions