GPT-5.3 Codex

GPT-5.3 Codex is a coding-optimized GPT model for repository edits, reviews, and agentic software work. It is part of the gpt-codex line, following GPT-5.2 Chat. It has a context window of 400000 tokens and a maximum output of 128000 tokens. It accepts image, pdf, and text inputs and returns text. It supports reasoning mode and tool calling. The weights are not open. Pricing starts at $1.40 per million input tokens and $11.20 per million output tokens, with cached input at $0.144 per million tokens.

Against GPT-5.2 Chat: window 128K to 400K, input $0.25 to $1.40, output $2.00 to $11.20, cached input $0.12 to $0.144.

Range across the hosts that serve it: up to $3.50 per 1M input tokens.

• ·400k token context window for large codebases
• ·128k token max output for long generations
• ·Reasoning mode for complex software tasks
• ·Tool calling for agentic workflows
• ·Accepts image, pdf, and text inputs

• Reach for it for repository edits and code reviews
• Reach for it for agentic software workflows
• Reach for it for analyzing large codebases with 400k context
• Reach for it for generating long code outputs up to 128k tokens

25 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 max output?

Context window is 400000 tokens, max output is 128000 tokens.

What are the pricing details?

Input from $1.40 per million tokens, output from $11.20 per million tokens, cached input from $0.144 per million tokens.

Does it support image inputs?

Yes, it accepts image, pdf, and text inputs, and returns text.