# OpenAI GPT-5.2-Codex

As of October 21, 2026, OpenAI GPT-5.2-Codex ranks #55 of 213 models overall with 28.5 on the Artificial Analysis Intelligence Index. Its strongest category is reasoning, where it ranks #40 of 203. OpenAI GPT-5.2-Codex costs $1.75 per million input tokens and $14.00 per million output tokens. It was released on January 14, 2026.

Page: https://llmbenchmark.ai/models/gpt-5-2-codex
Data from October 21, 2026. Every figure below names the board it was read from and the day.

- Maker: OpenAI
- Released: January 14, 2026
- Price per million input tokens: $1.75
- Price per million output tokens: $14.00
- Context window: 400K tokens
- Overall rank: 55 of 213, by the Artificial Analysis Intelligence Index
- Artificial Analysis Intelligence Index: 28.5
- Vibe Coding Index: 55.1 (from 2 of 4 coding benchmarks)

Benchmark results (22 of 190):

- SWE-bench Verified: 72.8% (source: SWE-bench (swebench.com), read October 21, 2026)
- SWE-bench Multilingual: 66.3% (source: SWE-bench (swebench.com), read October 21, 2026)
- LMArena WebDev: 1339 Elo (source: LMArena WebDev, read October 21, 2026)
- AA Intelligence Index: 28.5 (source: Artificial Analysis (via OpenRouter), read October 21, 2026)
- Design Arena: agents building web apps: 1051 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building full-stack apps: 993 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building mobile apps: 1090 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building Android apps: 1176 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building Godot games: 1142 Elo (source: Design Arena, read October 21, 2026)
- LiveCodeBench (Vals AI): 88.0% (source: Vals AI boards, read October 21, 2026)
- SWE-bench Verified (Vals AI): 72.4% (source: Vals AI boards, read October 21, 2026)
- Vibe Code Bench (Vals AI): 37.9% (source: Vals AI boards, read October 21, 2026)
- GPQA Diamond (Artificial Analysis run): 89.9% (source: Artificial Analysis, read October 21, 2026)
- Humanity's Last Exam (Artificial Analysis run): 35.7% (source: Artificial Analysis, read October 21, 2026)
- Terminal-Bench Hard (Artificial Analysis run): 37.1% (source: Artificial Analysis, read October 21, 2026)
- IFBench: 77.6% (source: Artificial Analysis, read October 21, 2026)
- Long context reasoning (AA-LCR): 82.3% (source: Artificial Analysis, read October 21, 2026)
- τ²-Bench Telecom: 92.1% (source: Artificial Analysis, read October 21, 2026)
- Terminal-Bench: 66.5% (source: Epoch AI benchmarking hub, read October 21, 2026)
- ALE-Bench: 1299.9 (source: ALE-Bench (Sakana AI), read October 21, 2026)
- LMArena WebDev: React: 1329 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena WebDev: HTML: 1348 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)