DeepSeek V3.1 Terminus vs GPT-5.2-Codex
DeepSeek V3.1 Terminus is cheaper on output tokens, while GPT-5.2-Codex offers a larger context window. Choose DeepSeek V3.1 Terminus or GPT-5.2-Codex based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | DeepSeek V3.1 Terminus | GPT-5.2-Codex |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Input / 1M tokens | $0.27 | $1.75 |
| Output / 1M tokens | $1.00 | $14.00 |
| Context window | 164K | 400K |
| Parameters | — | — |
| Open weights | — | No |
| Released | — | Dec 2025 |
| LMArena (Chatbot Arena) Elo score | 1419 | — |
| LiveBench score | — | 74.3 |
| SWE-bench Multilingual score | — | 66.3 |
| SWE-bench Verified score | — | 72.8 |
| LMArena WebDev Elo score | — | 1338 |
FAQ
Is DeepSeek V3.1 Terminus or GPT-5.2-Codex cheaper?
DeepSeek V3.1 Terminus is cheaper on output tokens ($1.00 vs $14.00 per 1M).
Which has the larger context window, DeepSeek V3.1 Terminus or GPT-5.2-Codex?
GPT-5.2-Codex has the larger context window (400K tokens).
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.