How to Use GPT-oss-120B in Amp
Use GPT-oss-120B for coding, math, and tasks that combine reasoning with tool calls. Its adjustable reasoning effort lets you choose how much analysis a task needs.
Prompts to try with GPT-oss-120B in Amp
Copy an example into Amp, or ask in your own words.
Use GPT-oss-120B with high reasoning effort to investigate the failing tests in this project. Read the relevant code, reproduce the failure, make the smallest justified fix, and run the affected tests.
Use GPT-oss-120B to review the current diff against the project's requirements. Report only actionable bugs, with a concrete failure scenario and the test that would expose each one.
Use GPT-oss-120B to compare this project's request validation code with its documented contract. Identify mismatches, add focused regression coverage, and explain any behavior that still needs a product decision.
Use GPT-oss-120B with high reasoning effort to check the numerical calculation in this project. Derive expected results for boundary cases, compare them with the implementation, and fix confirmed discrepancies.
Use GPT-oss-120B to inspect the project's text extraction pipeline. Check missing values, malformed input, and output validation; propose a focused patch with representative examples.
How to set up GPT-oss-120B 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-oss-120B
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": "openrouter", "model": "openai/gpt-oss-120b", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Amp finds GPT-oss-120B on Spiral, shows the price and runs it.
GPT-oss-120B providers and prices
Spiral runs GPT-oss-120B through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.001922 per run at the default inputs. $0.117 per 1M input tokens, $0.468 per 1M output tokens.
Settings you can ask for
Ask Amp for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Seed | a whole number | ||
| Min p | 0 to 1 | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Logprobs | on or off | ||
| Max tokens | 1 to 65536 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Reasoning effort | low, medium, or high | medium | How much the model reasons. |
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of OpenAI: gpt-oss-120b is on its tool page: open its card above.
What can you do with GPT-oss-120B in Amp?
GPT-oss-120B can help turn failing tests and code changes into fixes you can verify. Give it the relevant project files, expected behavior, and test output so it can propose a patch and check the result with available tools.
Debug a failing test
Provide the failing test, error output, and relevant code. Ask for the smallest fix and a regression test that reproduces the failure.
Review a code change
Compare a patch against its requirements. Ask for concrete failure cases, affected code paths, and tests that would confirm each finding.
Check a mathematical solution
Supply the problem and proposed answer. Ask for an independent check, assumptions, and a concise explanation of any disagreement.
Extract structured records
Turn supplied text into JSON records with defined keys. Include examples and require missing values to remain null rather than be guessed.
Answer questions from retrieved sources
Give the model relevant search results or reference passages. Request an answer with supporting excerpts and an explicit list of unanswered questions.
Compare technical proposals
Provide competing designs and acceptance criteria. Ask for a comparison of assumptions, contradictions, and unresolved implementation choices.
Investigate numerical discrepancies
Supply reported totals and underlying rows. With an available Python tool, ask the model to recompute the figures and identify where they diverge.
More on GPT-oss-120B and Amp
Overview
GPT-oss-120B in other agents
Questions about GPT-oss-120B in Amp
Can Amp use GPT-oss-120B?
Yes. Add Spiral to Amp as an MCP server, and Amp can find and run GPT-oss-120B like any built-in tool.
How much does GPT-oss-120B cost in Amp?
Up to $0.003202 per run at the default inputs. $0.195 per 1M input tokens, $0.78 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 OpenRouter account to use GPT-oss-120B in Amp?
No. Spiral runs GPT-oss-120B on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up GPT-oss-120B in Amp?
About two minutes: add the server, then sign in, then check it.
Does Amp wait while GPT-oss-120B runs?
GPT-oss-120B runs in the background. Amp starts the run, checks on it and gives you the text, json when it's ready.
What is GPT-oss-120B good at?
Coding, mathematical problem solving, and tasks that require tool calls are supported by OpenAI's evaluations. Practical starting points include debugging code, checking calculations, and turning retrieved information into structured records.
How much text can GPT-oss-120B handle?
OpenAI documents a 131,072-token context window. For long requests, include the relevant passages and specify the output you need, such as a short comparison or a table of discrepancies.
Can GPT-oss-120B analyze pictures or recordings?
GPT-oss-120B does not accept images, audio, or video. Supply a text transcript or extracted text for analysis.
When should I increase reasoning effort?
Try low effort for straightforward dialogue and high effort for difficult debugging or problems with several constraints. Compare the results on representative tasks before making high effort your default.
Can GPT-oss-120B work in languages other than English?
OpenAI evaluated it in 14 languages, including Spanish, French, German, and Japanese. Scores varied by language, so check examples in the language and subject you plan to use.
Start using GPT-oss-120B 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


