How to Use GPT-5.3-Codex in Amp
GPT-5.3-Codex writes, reviews, and debugs code. Its scope also covers product requirements, tests, and monitoring, making it useful for work that continues beyond the code change.
Prompts to try with GPT-5.3-Codex in Amp
Copy an example into Amp, or ask in your own words.
Use GPT-5.3-Codex with medium reasoning effort to investigate this project's failing tests. Reproduce the first failure, fix its cause, and run the relevant checks. Report the change and any remaining failures.
Use GPT-5.3-Codex with high reasoning effort to review the current diff. Focus on bugs and backward compatibility. For each finding, identify the location, a failure scenario, and the test that would catch it.
Use GPT-5.3-Codex to add pagination to this project's results screen. Follow existing components, preserve filtering behavior, and cover empty results and the final page. Run the relevant tests.
Use GPT-5.3-Codex to trace how this project's settings are saved. Fix the case where the interface reports success but a refresh loses the change. Add a regression test and verify persistence.
Use GPT-5.3-Codex to compare the feature requirements with the implementation and tests in this project. Identify missing acceptance criteria, implement the clearly specified behavior, and report unresolved product decisions.
How to set up GPT-5.3-Codex in Amp
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add the server
No header and no client ID or secret: Amp registers itself with Spiral.
amp mcp add spiral https://api.spiral.ai/mcp
- Sign in
Start amp. Amp opens Spiral's sign-in page in your browser: sign in, pick the organization and click Authorize. No API key.
- Check it
amp mcp doctor
Use the Spiral skill
- Send Amp this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- Install and sign in
Installs the Spiral CLI, signs you in and adds its skill to your coding agents.
npx -y spiral-cli@latest init
Call the Spiral API
- Create an API key
In the console, under API keys.
- Run GPT-5.3-Codex
One request; the result comes back by polling or a webhook.
curl https://api.spiral.ai/v1/runs \ -H "Authorization: Bearer $SPIRAL_API_KEY" \ -H "Content-Type: application/json" \ -H "Idempotency-Key: $(uuidgen)" \ -d '{ "provider": "openai", "model": "openai/gpt-5.3-codex", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Amp finds GPT-5.3-Codex on Spiral, shows the price and runs it.
GPT-5.3-Codex providers and prices
Spiral runs GPT-5.3-Codex through OpenAI. Open a card for today's price, with every fee included.
Up to $0.074628 per run at the default inputs. $2.275 per 1M input tokens, $0.2275 per 1M cached input tokens, $18.20 per 1M output tokens.
Settings you can ask for
Ask Amp for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Max tokens | 16 to 128000 | The most output tokens to generate. | |
| Reasoning effort | low, medium, high, or xhigh | How much the model reasons. | |
| Parallel tool calls | on or off |
Every API field of GPT-5.3-Codex is on its tool page: open its card above.
What can you do with GPT-5.3-Codex in Amp?
Turn a reproducible bug or feature request into code, tests, and a review of the result with GPT-5.3-Codex. Point the agent to relevant project files and define the behavior it should verify.
Reproduce and repair a bug
Give GPT-5.3-Codex the failure, expected behavior, and relevant code. Request a focused fix and a regression test.
Review a proposed change
Ask GPT-5.3-Codex to inspect a diff for behavior changes and explain each finding with a concrete failure scenario.
Design regression tests
Have GPT-5.3-Codex turn a bug report or requirement into tests with explicit inputs and expected results.
Implement a web feature
Ask GPT-5.3-Codex to write a feature with defined loading, empty, success, and error behavior.
Draft product requirements
Use GPT-5.3-Codex to turn a feature brief into requirements, acceptance criteria, and unresolved product decisions.
Triage monitoring alerts
Have GPT-5.3-Codex compare alert details, logs, and recent changes, then propose checks that distinguish possible causes.
Edit product copy
Ask GPT-5.3-Codex to revise interface messages against the specified behavior and flag claims the requirements do not support.
More on GPT-5.3-Codex and Amp
Overview
GPT-5.3-Codex in other agents
More Amp guides
Questions about GPT-5.3-Codex in Amp
Can Amp use GPT-5.3-Codex?
Yes. Add Spiral to Amp as an MCP server, and Amp can find and run GPT-5.3-Codex like any built-in tool.
How much does GPT-5.3-Codex cost in Amp?
Up to $0.074628 per run at the default inputs. $2.275 per 1M input tokens, $0.2275 per 1M cached input tokens, $18.20 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Amp plan is separate.
Do I need an OpenAI account to use GPT-5.3-Codex in Amp?
No. Spiral runs GPT-5.3-Codex on its own OpenAI account. You sign in to Spiral once and never paste a key.
How long does it take to set up GPT-5.3-Codex in Amp?
About two minutes: add the server, then sign in, then check it.
Does Amp wait while GPT-5.3-Codex runs?
GPT-5.3-Codex runs in the background. Amp starts the run, checks on it and gives you the text, json when it's ready.
What is GPT-5.3-Codex useful for?
Use it to write, review, or debug code and to handle related work such as test design, product requirements, and monitoring analysis. Give it a concrete task and a clear definition of done.
How much text can GPT-5.3-Codex handle?
OpenAI lists a 400,000-token context window, a maximum of 272,000 input tokens, and up to 128,000 output tokens. Treat these as ceilings when deciding how much material to include.
Which reasoning effort should I use?
Start with medium for interactive coding. Try high or extra high for difficult debugging or other tasks where the first attempt needs deeper reasoning.
What should I include in a coding prompt?
Describe the expected behavior, provide the relevant code or reproduction steps, and state the constraints. Explain how to verify success, such as a failing test that should pass after the fix.
Is GPT-5.3-Codex intended for casual conversation?
OpenAI says it is not intended for conversational use. In a chat app, give it focused technical tasks such as reviewing pasted code, explaining a failure, or drafting test cases.
Start using GPT-5.3-Codex in Amp
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Amp:
set up https://spiral.ai/SKILL.md


