GPT-5.2 Pro

GPT-5.2 Pro positions itself among the most capable reasoning models in the current landscape. Its benchmark profile reveals particular strength in graduate-level reasoning tasks, where it achieves a score of 93.20 on GPQA Diamond, and advanced mathematical problem-solving, reaching 31.30 on FrontierMath Tier 4. Solid performance on high-level evaluations (HLE at 50) rounds out a profile suggesting a model engineered for multi-step analytical challenges and nuanced, knowledge-intensive tasks that demand deep comprehension alongside persistent accuracy.

The model integrates into OpenAI's broader developer ecosystem, supporting structured outputs, function calling, and tool-augmented workflows. This positions GPT-5.2 Pro as well-suited for applications requiring reliable API integration alongside advanced reasoning—spanning complex code generation, multi-step agent tasks, and research-adjacent workloads where benchmark-leading performance on mathematical and analytical benchmarks translates to practical utility across demanding domains.

DevPass (LLM Gateway)gpt-5.2-progpt-pro

Open provider page Provider docs

Quick Info

Provider: DevPass (LLM Gateway)
Model key: gpt-5.2-pro
Release date: Dec 11, 2025
Last updated: Dec 11, 2025
Knowledge cutoff: 2025-08-31
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $21.00
Output token cost: $168.00

Limits

Input tokens: 272,000 tokens
Output tokens: 128,000 tokens
Context window: 400,000 tokens

Transparent token rates

Compare GPT-5.2 Pro pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-5.2 Pro

OpenAI

Official sourceAnnouncement

OpenAI introduced GPT-5.2 on December 11, 2025, as its most capable model series for professional knowledge work and long-running agents. The launch explicitly names GPT-5.2 Instant, Thinking, and Pro as three tiers rolling out in ChatGPT starting with paid plans, with all variants also available in the API on the same day. The release targets spreadsheets, presentations, code, image perception, long context, tool use, and complex multi-step projects. GPT-5.2 Thinking sets new state-of-the-art results on benchmarks including GDPval (70.9% wins or ties vs GPT-5's 38.8%), SWE-Bench Verified (80.0%), GPQA Diamond (92.4%), AIME 2025 (100%), and ARC-AGI-2 (52.9% vs 17.6%). Partners such as Notion, Box, Shopify, Harvey, Zoom, Databricks, Cognition, JetBrains, and Augment Code report state-of-the-art long-horizon reasoning, agentic coding, and tool-calling performance.