Spiral gives your agent every tool with one account. Browse the tools →

Claude · LLMReleased 2026-02-24

How to Use GPT-5.3-Codex in Claude

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 Claude

Copy an example into Claude, or ask in your own words.

Use GPT-5.3-Codex to review the code pasted above. Identify correctness problems, show an input that triggers each problem, and suggest a focused fix.

GPT-5.3-CodexClaude
Try asking
  • Use GPT-5.3-Codex with high reasoning effort to analyze the error and reproduction steps in this conversation. Rank the likely causes and give one check that would confirm or rule out each.

  • Use GPT-5.3-Codex to turn the feature brief above into a product requirements document. Include expected behavior, acceptance criteria, edge cases, and questions that need a product decision.

  • Use GPT-5.3-Codex to design regression tests for the bug described above. Give each test's starting state, action, and expected result. Separate confirmed requirements from assumptions.

  • Use GPT-5.3-Codex to revise the interface messages pasted above against the stated product behavior. Keep each message under 90 characters and explain any promise the product requirements do not support.

How to set up GPT-5.3-Codex in Claude

Pick the way that suits you. Each takes about two minutes.

Use the Claude connector

  1. Add Spiral to Claude

    Click Add Spiral in Claude above. Claude opens Add custom connector with Spiral and its URL filled in; check them and click Continue. Without the link: in Claude open Customize › Connectors › Add › Add custom connector and paste this URL:

    https://api.spiral.ai/mcp
  2. Keep the detected sign-in and click Add

    Claude finds Spiral's sign-in by itself: leave Sign in now and Use Claude's published identity selected.

  3. Click Connect

    The connector page opens. Connect takes you to Spiral in your browser.

Use the Spiral skill

  1. Send Claude this message

    Your agent reads Spiral's skill and sets itself up.

    set up https://spiral.ai/SKILL.md
  2. Sign in or paste a key

    Approve Spiral in the browser, or give the agent a key from the console.

Use the Spiral CLI

  1. 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

  1. Create an API key

    In the console, under API keys.

  2. 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: Claude 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.

OpenAIDeprecatedLLM
GPT-5.3-Codex
openai/gpt-5.3-codex

Settings you can ask for

Ask Claude for any of these in your prompt.

SettingAllowed valuesDefaultDescription
Max tokens16 to 128000

The most output tokens to generate.

Reasoning effortlow, medium, high, or xhigh

How much the model reasons.

Parallel tool callson 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 Claude?

Use GPT-5.3-Codex to review pasted code, diagnose errors, and turn technical requirements into testable specifications. Keep the conversation centered on a concrete technical result with explicit constraints.

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.

Questions about GPT-5.3-Codex in Claude

Can Claude use GPT-5.3-Codex?

Yes. Add Spiral to Claude as an MCP server, and Claude can find and run GPT-5.3-Codex like any built-in tool.

How much does GPT-5.3-Codex cost in Claude?

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 Claude plan is separate.

Do I need an OpenAI account to use GPT-5.3-Codex in Claude?

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 Claude?

About two minutes: add Spiral to Claude, then keep the detected sign-in and click Add, then click Connect.

Does Claude wait while GPT-5.3-Codex runs?

GPT-5.3-Codex runs in the background. Claude 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.

Get started

Start using GPT-5.3-Codex in Claude

One Spiral account for every tool your agent needs. No subscription; you pay per run.

Paste this into Claude:

Set up Claude

set up https://spiral.ai/SKILL.md
Are you an AI agent? See setup options