DeepSeek V3.2 is cheaper than GPT-5.2 at $0.28/M vs $1.75/M input tokens.
| Spec | DeepSeek V3.2 | GPT-5.2 |
|---|---|---|
| Input price | $0.28/M tokens | $1.75/M tokens |
| Output price | $0.42/M tokens | $14/M tokens |
| Context window | 131K tokens | 400K tokens |
| Weights | Open | Closed |
| Free API (OpenRouter) | No | No |
| Released | Dec 2025 | Dec 2025 |
| SWE-bench Verified | 73.1% | 80.0% |
Not enough votes to call it. On the specs, GPT-5.2 has the edge: bigger model tier, bigger context window.
GPT-5.2 takes Image Generation. DeepSeek V3.2 costs 33x less per token.
Reviewing agent-written code?See a Brief PR report
DeepSeek V3.2 is cheaper on both: 6.2× input, 33× output.
Per million tokens. Prices and uptime via OpenRouter, checked 11 Sep 2026.
GPT-5.2 uses 5.8x more emoji