Model

GPT-5.2

CLOSED

OpenAI · GPT-5.2 family · released Dec 11, 2025

Adds GPT-5.2 Pro tier and new xhigh reasoning effort; SOTA at launch on GDPval, ARC-AGI-2, and long-context MRCRv2.

ReasoningCodingVisionFunction callingTool useAgentic

2495.0

Elo · rank #18

Parameters

Undisclosed

Active params

—

Context

400K tokens

Architecture

Undisclosed architecture (presumed dense transformer, reasoning-tuned); shipped as Instant/Thinking/Pro tiers

License

Proprietary

Languages

—

API price (in/out)

$1.75 / $14

Modalities

text · vision

Benchmark results

Bar shows position within the tracked field; marker = field best

AIMEMath100.0%#1

best: this model · 100.0%

ARC-AGI-1Reasoning86.2%#5

best: GPT-5.6 · 96.5%

ARC-AGI-2Reasoning52.9%#11

best: GPT-5.6 · 92.5%

FrontierMath (Tiers 1-3)Math40.3%#8

best: GPT-5.6 · 89.0%

GDPvalAgents70.9%#5

best: Seed 2.1 Pro · 87.9%

GPQA DiamondReasoning92.4%#14

best: GPT-5.6 · 94.6%

Humanity's Last ExamReasoning34.5%#37

best: Claude Sonnet 5 · 57.4%

LiveCodeBenchCoding88.9%#10

best: DeepSeek-V4-Pro (Think Max) · 93.5%

SWE-bench ProCoding55.6%#23

best: Claude Fable 5 · 80.0%

SWE-bench VerifiedCoding80.0%#18

best: Claude Fable 5 · 95.0%

τ²-Bench TelecomAgents98.7%#4

best: Claude Opus 4.6 · 99.3%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$1.75

Output / M tok

$14

GPT-5.2 family

Elo progression across releases

GPT-5.2Dec 20252495.0

Resources

DOCSplatform.openai.com/docs/models/gpt-5.2 APIAvailable via OpenAI API

API price $1.75/$14 · each benchmark row carries its own source badge (see methodology)