Claude code execution

by AnthropicCode execution

Code execution is Anthropic's server tool that gives Claude a sandboxed container to run Bash commands and Python and to create and edit files while it answers. The container has 5 GiB of memory and no internet access, so only its preinstalled libraries are there, and you can reuse it across requests for up to 30 days to keep the files Claude made. On Opper, Claude gets it on the Messages endpoint, passed through as Anthropic defines it. It runs on Anthropic's own route; Claude served from Bedrock or Vertex doesn't have it.

UseClaude code executionwithin

Add Claude code execution on Opper

Set up with agents

Paste this into your coding agent (Claude Code, Cursor, Codex, and more) and it will set up and build with Opper for you.

Get your API key

Key info

Tool

code_execution

Endpoint

Messages

Routes

Anthropic

In your code

1curl https://api.opper.ai/v3/compat/v1/messages \
2  -H "Authorization: Bearer $OPPER_API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{"model": "anthropic/claude-haiku-5-5", "max_tokens": 1024,
5       "messages": [{"role": "user", "content": "Sum the first 50 primes and show your working."}],
6       "tools": [{"type": "code_execution_20260521", "name": "code_execution"}]}'

Use Claude code execution with these models

Step-by-step setup for Claude code execution on each of these models through Opper.

Claude Haiku 5.5 Claude Sonnet 5.5 Claude Opus 5.5 Claude Fable 5.1 Claude Opus 5 Claude Sonnet 5 Claude Fable 5 Claude Opus 4.8 See all 700+ models

More tools on Opper

OpenAI code interpreter

GPT models write and run Python in a sandbox, on the Responses endpoint.

Codecode_interpreter

Gemini code execution

Gemini runs code to work out its answer, on generateContent.

CodecodeExecution

Claude computer use

Claude drives a desktop through screenshots, clicks and typing in your agent loop.

Codecomputer_toolset