OpenAI

GPT-5.2

OpenAI flagship. 74.9% SWE-bench. Best vision model. Thinking mode is slow but powerful.

7.4

8 reviews

Input Cost$1.75/M

Output Cost$14.00/M

Context Window400K tokens

Capabilities

VisionFunctionsTools

Community Ratings

Vision Quality

8.8

Tool Use

8.4

Code Quality

8.2

Instruction Following

8.1

Context Handling

8.1

Consistency

7.0

Value for Money

6.3

Speed Feel

5.1

Used this model? Share your experience.

Write a Review

Reviews (8)

@dev_reviewer_8Editorial

7.3

Tool Use1mo ago

Enhanced tool-calling and agentic workflows. GitHub Copilot integration is solid.

@dev_reviewer_6Editorial

7.1

Tool Use1mo ago

Enhanced tool-calling and agentic workflows. GitHub Copilot integration is solid.

@dev_reviewer_7Editorial

7.4

Tool Use1mo ago

Enhanced tool-calling and agentic workflows. GitHub Copilot integration is solid.

@dev_reviewer_4Editorial

7.7

Vision1mo ago

Strongest vision model at release. Error rates halved on chart reasoning and UI understanding.

@dev_reviewer_5Editorial

7.2

Vision1mo ago

Strongest vision model at release. Error rates halved on chart reasoning and UI understanding.

@dev_reviewer_3Editorial

7.2

Vision1mo ago

Strongest vision model at release. Error rates halved on chart reasoning and UI understanding.

@dev_reviewer_1Editorial

7.4

Coding1mo ago

Incredibly impressive but too slow. Thinking mode is painful. 15-20% fewer mid-chain errors though.

@dev_reviewer_2Editorial

7.5

Coding1mo ago

Incredibly impressive but too slow. Thinking mode is painful. 15-20% fewer mid-chain errors though.