GPT-5.2 is slightly cheaper per million tokens (blended). GPT-5.4 leads quality index by 4.2 points (79.3 vs 75.1). GPT-5.4 has the larger context window (1,050,000 vs 400,000 tokens, 2.6×).

GPT-5.2 vs GPT-5.4

Live catalog fields. Best-in-row is highlighted.

GPT-5.2

OpenAI · openai__gpt-5.2

GPT-5.4

OpenAI · openai__gpt-5.4

Identity

Fieldopenai__gpt-5.2openai__gpt-5.4
ProviderOpenAIOpenAI
Slugopenai__gpt-5.2openai__gpt-5.4
Statuslivelive
Open weightsNoNo
License
HuggingFace

Quality

Fieldopenai__gpt-5.2openai__gpt-5.4
Quality index (0–100)75.179.3

Cost

Fieldopenai__gpt-5.2openai__gpt-5.4
Input $ / 1M tokens$1.75$2.50
Cached input $ / 1M$0.17$0.25
Output $ / 1M tokens$14.00$15.00

Context

Fieldopenai__gpt-5.2openai__gpt-5.4
Max context tokens400,0001,050,000
Max output tokens128,000128,000

Modalities

Fieldopenai__gpt-5.2openai__gpt-5.4
Inputfile, image, texttext, image, file
Outputtexttext