Model

GPT-5.5 (xHigh)

CLOSED

OpenAI · GPT-5.5 family · released Apr 23, 2026

Highest reasoning-effort sibling of the base 'GPT-5.5' entry ('reserved for the hardest quality-first workloads' per OpenAI's own guidance). Same weights/pricing/context as the default and High siblings; distinguished by reasoning.effort=xhigh.

ReasoningCodingVisionFunction callingTool useAgentic

2725.9

Elo · rank #6

Parameters

Undisclosed

Active params

—

Context

1.05M tokens

Architecture

Undisclosed architecture (presumed dense transformer, reasoning-tuned); codename "Spud"; 'xhigh' (maximum) reasoning-effort configuration

License

Proprietary (OpenAI API Terms)

Languages

—

API price (in/out)

$5 / $30

Modalities

text · vision

Benchmark results

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

Agents' Last ExamAgents47.9%#3

best: GPT-5.6 · 52.7%

FrontierMath (Tiers 1-3)Math51.7%#4

best: GPT-5.6 · 89.0%

FrontierMath (Tier 4)Math35.4%#4

best: GPT-5.6 · 83.0%

GDPval-AAAgents1500#14

best: Claude Fable 5 · 1932

GPQA DiamondReasoning93.5%#7

best: GPT-5.6 · 94.6%

Humanity's Last ExamReasoning44.0%#15

best: Claude Sonnet 5 · 57.4%

SWE-bench VerifiedCoding80.6%#14

best: Claude Fable 5 · 95.0%

Terminal-Bench 2.0Coding84.0%#4

best: GPT-5.6 · 88.8%

Run it locally

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

Input / M tok

Output / M tok

$30

GPT-5.5 family

Elo progression across releases

GPT-5.5Apr 20262680.9 GPT-5.5 (High)Apr 20262563.6 GPT-5.5 (xHigh)Apr 20262725.9

Resources

DOCShttps://developers.openai.com/api/docs/models/gpt-5.5 APIAvailable via OpenAI API

API price $5/$30 · each benchmark row carries its own source badge (see methodology)