OpenAI: GPT-5.3-Codex
gpt-5.3-codex- Input / Output-30%
- $1.75 / $14.00$1.23 / $9.80
- Context
- 400K
- Max output
- 128K
- Modalities
- VisionChat
- Capabilities
- Tool usePrompt Cache
About GPT-5.3-Codex
GPT-5.3-Codex is OpenAI's agentic coding model, called the most capable of its Codex line at release and tuned for Codex-style environments. It handles long software tasks with reasoning effort from low to xhigh, takes text and image input, and uses tools like hosted shell and skills.
Where it works well
- Tuned for agentic coding in Codex-style environments rather than general chat.
- Reasoning effort ranges from low to xhigh for harder software tasks.
- Hosted shell and skills are listed among its supported tools.
- Image input lets it work from screenshots of errors or interfaces.
When to choose another model
- It is built for coding agents, so quick conversational turns and light drafting are better handled by a general GPT model.
- It has no none effort, so very quick turns cost more reasoning than on general models.
- Output is text only, so it cannot produce images or audio.
Getting started
Create API key
Create a key in Console, then use it with every model on the platform.
Send your first request
Copy the example for your language and run it against the endpoint.
Text to textResponses APIPOST/v1/responsesAPI format:curl https://api.tokenlab.sh/v1/responses \ -H "Content-Type: application/json" \ -H "Authorization: Bearer sk-xxx" \ -d '{ "model": "gpt-5.3-codex", "input": "Hello!" }'
Pricing
The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.
Default spec
per 1M tokens- Official price
- Input $1.75 / Output $14.00 / Cache read $0.175
- Verified price
- Input $1.23 / Output $9.80 / Cache read $0.1225
- Discount
- -30%
Cache read
- Official price
- $0.175
- Verified price
- $0.1225
- Discount
- -30%
Charged per call, only when the model uses the tool.
Web search
- Official price
- $0.01/search
- Verified price
- $0.007/search
- Discount
- -30%
| Official priceper 1M tokens | Verified priceper 1M tokens | Discount | |
|---|---|---|---|
| Default spec | Input $1.75 / Output $14.00 / Cache read $0.175 | Input $1.23 / Output $9.80 / Cache read $0.1225 | -30% |
| Prompt cache pricing | |||
| Cache read | $0.175 | $0.1225 | -30% |
| Tool feesCharged per call, only when the model uses the tool. | |||
| Web search | $0.01/search | $0.007/search | -30% |
Usage & activity
Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.
Usage & availability
Last 24 hours- Requests
- Success rate
- P95 latency
- Total tokens
Data is based on aggregate user requests, excluding status checks.
Open in Console
Open GPT-5.3-Codex in Console with a prompt ready to edit or send.
Help me try gpt-5.3-codex with a short message at /v1/responses. Show the reply, latency, and cost.
Use cases
Best for- Vision
Repository changes
Read a codebase, implement a feature across files, run the tests, and iterate on failures.
Code review
Inspect a diff for defects, explain each concern, and propose a fix that the author can apply directly.
Bug reproduction
Take a screenshot and log of a failure, then write a minimal reproducing test.
Migration work
Update API usage across a project and confirm the build still passes.
Prompt examples
Implement pagination in the orders endpoint, update the client, and run the tests until they pass.
Review this diff for concurrency bugs and write a failing test for each one.
Attached are a screenshot of the error page and the server log. Find the cause and patch it.
This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.
FAQ
What is OpenAI's GPT-5.3-Codex model?
GPT-5.3-Codex is an OpenAI model specialized for agentic coding. OpenAI describes it as its most capable agentic coding model at the time of release. It is meant to run inside coding agents and editors.
What reasoning levels does GPT-5.3-Codex support?
OpenAI lists effort levels from low up to xhigh. Use the higher levels for large refactors and subtle bugs, and the lower ones for small mechanical edits.
Does GPT-5.3-Codex accept images?
Yes. Text and image input are supported, which helps when the starting point is a screenshot of an error or interface. The output is text.
Does GPT-5.3-Codex work with tools?
Yes. OpenAI lists hosted shell and skills among its supported tools, so an agent can run shell commands and reuse packaged skills while it works through a repository task. That is the setting its Codex-style tuning is aimed at.
How does GPT-5.3-Codex differ from GPT-5.2?
GPT-5.3-Codex is specialized for software engineering agents, while GPT-5.2 is a general professional-work flagship that also offers a none effort option for quick turns. Choose by workload: repository agents or general knowledge work.
How much does GPT-5.3-Codex cost?
On TokenLab, GPT-5.3-Codex costs Input $1.23 / Output $9.80 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.
What are the context window and output limit of GPT-5.3-Codex?
GPT-5.3-Codex accepts up to 400,000 tokens of context and returns up to 128,000 tokens in one response.
Which endpoint should GPT-5.3-Codex use?
Use https://api.tokenlab.sh/v1/responses for GPT-5.3-Codex. The request example below shows the matching code shape.
Which operations does GPT-5.3-Codex support?
GPT-5.3-Codex supports Text to text. Select an operation above to see its endpoint and request example.
Compare GPT-5.3-Codex
Sources
Reviewed Oct 2, 2026
More from OpenAI