Code execution
Give the model a sandbox and it writes and runs code to answer: maths, data analysis, charts and file transforms, with the output folded into its reply. OpenAI's code interpreter, Gemini's code execution and Claude's code execution all run through Opper on the same key, passed through as each provider defines them. Claude's runs on Anthropic's own route, not on Bedrock or Vertex.
Add 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.
In your code
curl https://api.opper.ai/v3/compat/responses \-H "Authorization: Bearer $OPPER_API_KEY" \-H "Content-Type: application/json" \-d '{"model": "openai/gpt-6-sol","input": "Sum the first 50 primes and show your working.","tools": [{"type": "code_interpreter", "container": {"type": "auto"}}]}'
Code execution tools on Opper
Each one has its own page with setup, code and the models it works with.
OpenAI code interpreter
GPT models write and run Python in a sandbox, on the Responses endpoint.
code_interpreterGemini code execution
Gemini runs code to work out its answer, on generateContent.
codeExecutionClaude code execution
Claude runs code in Anthropic's sandbox, on Anthropic's own route.
code_execution- OpenAI:
code_interpreteron/v3/compat/responses - Google:
codeExecutionon/v3/compat/v1beta/models/{model}:generateContent - Anthropic:
code_execution (Anthropic's own route only)on/v3/compat/v1/messages
More tools on Opper
Web search
Let any model search the web mid-answer and cite its sources. Native search where the model has it, Jina, Searchbase, Exa or Linkup everywhere else.
opper:web_searchWeb fetch
Turn any URL into clean markdown a model can read. Runs on Jina in the EU.
/v3/tools/web/fetchRerank
Reorder search results by relevance before they reach the model. Cohere Rerank and BGE.
/v3/rerank