GPT-5.3 Codex Spark

GPT-5.3 Codex Spark is a coding-optimized GPT model from OpenAI for repository edits, reviews, and agentic software work. It is the first version of the gpt-codex-spark line in the registry. It has a 128000-token context window and a 32000-token output limit. It accepts image, pdf, and text inputs and returns text. It supports reasoning mode and tool calling. Weights are not open. Pricing starts at $1.75 per million input tokens and $14.00 per million output tokens, with cached input at $0.175 per million tokens.

First version of its line in the registry.

• ·128k token context window for large codebases
• ·32k token max output for long edits
• ·Supports reasoning mode and tool calling
• ·Accepts image, pdf, and text inputs
• ·Coding-optimized for repository edits and reviews

• Reach for it for editing large repositories
• Reach for it for code review tasks
• Reach for it for agentic software workflows
• Reach for it for analyzing code with images or PDFs

3 hosts serve this model at the price above · the maker's documentation

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

What is the context window and output limit?

The context window is 128000 tokens and the maximum output is 32000 tokens.

Does it support tool calling and reasoning mode?

Yes, it supports both reasoning mode and tool calling.

What is the pricing?

Input tokens cost from $1.75 per million, output tokens $14.00 per million, and cached input $0.175 per million.