모델 목록으로
OpenAI독점

GPT-5.2

OpenAI's flagship general-purpose model. It has significantly improved reasoning and coding performance from the previous generation GPT-5.1. It supports a 256K token context window and delivers stable high performance across a wide range of tasks.

파라미터

Undisclosed

컨텍스트

256K

라이선스

Proprietary

출시일

2026-04-20

일본어 처리 능력

High-Quality JP

Multilingual model with strong Japanese language processing capabilities.

API 가격

입력 가격 (1M 토큰당)

$1.25

출력 가격 (1M 토큰당)

$10

과금 모드: standard

강점

  • High versatility
  • 256K long-context support
  • Instant version for high-speed, low-cost use
  • Rich ecosystem

약점

  • Pro version is expensive
  • Not open-source
  • Japanese processing may lag behind specialized models

활용 사례

  • General-purpose text generation
  • Coding assistance
  • Long-document summarization and analysis
  • Chatbots

심층 분석

Arena Elo (Text)

1436

#21 of 117 on BenchLM provisional leaderboard

GPQA Diamond

92.4%

#6 in Knowledge category; vs Gemini 3 Pro: 91.9%

SWE-Bench Verified

80.0%

vs Claude Opus 4.5: 80.9% (effectively tied)

GDPval (Knowledge Work)

70.9%

First model to beat human experts on 44-occupation benchmark

Context Window

400K tokens

~100% coherence at 256K on MRCRv2

API Pricing

$1.75/$14 per 1M tokens

40% increase over GPT-5.1; 90% cached-input discount

강점

  • Best-in-class professional knowledge work — first model to outperform human experts across 44 occupations on GDPval (70.9%)
  • Outstanding long-context coherence, maintaining near-perfect accuracy up to 256K tokens on MRCRv2
  • Strong graduate-level science and reasoning: 92.4% GPQA Diamond, 100% AIME 2025, 52.9% ARC-AGI-2

약점

  • Weak agentic performance relative to peers — ranked #29 on BenchLM with only 61.0/100 on agentic benchmarks
  • Already superseded by GPT-5.4 and GPT-5.5 within months; GPT-5.2 Thinking retiring June 3, 2026
  • 40% price increase over GPT-5.1 ($1.75/$14 vs $1.25/$10) with diminishing marginal gains in some categories

경쟁사 비교

ModelArenaSWEGPQAPrice
Claude Opus 4.5~1430*80.9%87.0%$5/$25
Gemini 3 Pro~1440*91.9%$2/$12
GPT-5.4 (Successor)N/A92.8%$2.50/$15

GPT-5.2, released December 11, 2025, represented OpenAI's aggressive response to competitive pressure from Google's Gemini 3 Pro and Anthropic's Claude Opus 4.5 — internally referred to as a 'Code Red' effort. The model introduced a three-tier architecture (Instant, Thinking, Pro) with adaptive reasoning capabilities that dynamically allocate compute based on query complexity. It achieved several milestone benchmarks: the first model to exceed 90% on ARC-AGI-1 (Pro tier), a perfect 100% on AIME 2025, and 70.9% on GDPval — making it the first AI to outperform human industry professionals across 44 occupations on OpenAI's proprietary knowledge-work benchmark.

The model's strongest positioning was in professional knowledge work, scientific reasoning (GPQA Diamond: 92.4%), long-context analysis (near-perfect retrieval at 256K tokens), and coding (SWE-Bench Verified: 80.0%). It delivered a 30-38% reduction in error-containing responses versus GPT-5.1, making it significantly more reliable for production workflows. However, it showed notable weaknesses in agentic tasks (ranked #29 on BenchLM), and its multimodal capabilities lagged behind Gemini 3 Pro on comprehensive vision benchmarks. The model also came at a 40% price premium over its predecessor.

Within the GPT-5 family's rapid evolution, GPT-5.2 served as the 'reasoning milestone' that set new benchmarks before being progressively superseded by GPT-5.3 (March 2026), GPT-5.4 (March 2026, adding 1M context and computer use), and GPT-5.5 (April 2026). As of May 2026, GPT-5.2 Thinking is scheduled for retirement on June 3, 2026, with GPT-5.4+ as the recommended upgrade path. It remains available for budget-conscious users who need strong reasoning at lower cost than the newest frontier models.

분석 생성일: 2026-05-23