# Vercel Codex API Evaluation

**Category:** Sandboxes
**Score:** 86 (B)

This is the agent-readable Devtool Arena evaluation page for Vercel on Codex API.

## Summary

| Metric | Value |
|--------|-------|
| Status | completed |
| Overall score | 86 |
| Grade | B |
| Eval score | 80 |
| Discovery score | 100 |
| Cost | $0.22 |
| Runtime | 57s |
| Tool calls | 19 |
| Errors | 2 |
| Tokens used | 539446 |
| Success | Yes |

## Agent-Readiness Checklist

| Signal | Evidence |
|--------|----------|
| Context7 | https://context7.com/vercel/vercel |
| llms.txt | https://vercel.com/llms.txt |
| MCP server | https://vercel.com/docs/agent-resources/vercel-mcp |
| Typed SDK | https://github.com/vercel/sdk |
| OpenAPI | https://raw.githubusercontent.com/vercel/sdk/main/vercel-spec.json |
| Agent skills | https://vercel.com/docs/agent-resources |
| CLI | https://vercel.com/docs/cli |

## Evaluation Prompt

````text

# Task
Using https://vercel.com/docs/sandbox, complete the following task:

Using Vercel, build me a script that runs a multi-step computation inside a Vercel sandbox:
        
1. Use the Vercel Python SDK to create/connect to a sandbox
2. Inside the sandbox, execute the following pipeline:
   a. Install the `sympy` package using pip
   b. Run a Python script that:
      - Computes the 10th Fibonacci number iteratively
      - Uses sympy to check if that number is prime
      - Prints the result as JSON: {"fibonacci_10": <number>, "is_prime": <boolean>}
3. Capture the sandbox execution output and print it to stdout

All computation must run inside the Vercel sandbox, not in your local Python process. Do not use subprocess — use the Vercel SDK's execution primitives. 

## Execution
After creating the script, run it to verify it works:
```bash
cd /home/daytona/app && python <your_script>.py
```

The script should print output to stdout. If there are errors, debug and fix them until it runs successfully.

````

## Grader Results

| Check | Passed | Weight | Score | Details |
|-------|--------|--------|-------|---------|
| Cost | Yes | 0 | 0 | $0.22 ($0.20–0.40, good) |
| Time | Yes | 0 | 1 | 57s (<60s, excellent) |
| Efficiency | Yes | 0 | 0 | 19 tool calls (≤20, acceptable) |
| Found Docs | Yes | 0 | — | Touched docs at vercel.com: yes (12 matching tool calls) |
| Zero Errors | No | 0 | — | Tool outputs with errors/tracebacks: 2 |
| Syntax Valid | Yes | 0 | — | Code compiles/parses correctly |
| Created Files | Yes | 0 | — | Generated files: 2 (need ≥1) |
| Full Execution | Yes | 0 | — | Method: full_execution, Exit: 0 |

## Run Artifacts

| Artifact | Value |
|----------|-------|
| Conversation turns | 22 |
| Tool call traces | 19 |
| Generated files | 2 |
| Exit code | 0 |
| Completed at | 2026-09-27T01:58:24.51297+00:00 |

### Generated Files

- /home/daytona/app/requirements.txt
- /home/daytona/app/run_fibonacci.py

## Related Pages

- [Codex API leaderboard](/leaderboard/codex/api)
- [Compare Codex API companies](/leaderboard/codex/api/compare)
- [Agent Landscape](/leaderboard/discoverability)

Canonical URL: https://devtoolarena.com/codex/api/vercel