Parallel Tool Calls: Running Agent Tools Concurrently

When a request needs several independent lookups ("compare the weather in Delhi, Mumbai and Bengaluru"), the model can return multiple tool_use blocks in a single response. Running them one after another wastes time.

Execute in parallel

const toolUses = response.content.filter(
  (b): b is Anthropic.ToolUseBlock => b.type === 'tool_use'
);

const results: Anthropic.ToolResultBlockParam[] = await Promise.all(
  toolUses.map(async (tu) => {
    try {
      return { type: 'tool_result', tool_use_id: tu.id, content: await runTool(tu.name, tu.input) };
    } catch (err) {
      return { type: 'tool_result', tool_use_id: tu.id, content: String(err), is_error: true };
    }
  })
);

messages.push({ role: 'assistant', content: response.content });
messages.push({ role: 'user', content: results }); // ALL results in ONE message

Python with asyncio

import asyncio, json

async def run_all(tool_uses):
    async def one(tu):
        try:
            return {"type": "tool_result", "tool_use_id": tu.id,
                    "content": json.dumps(await run_tool(tu.name, tu.input))}
        except Exception as e:
            return {"type": "tool_result", "tool_use_id": tu.id,
                    "content": str(e), "is_error": True}
    return await asyncio.gather(*(one(tu) for tu in tool_uses))

Important rules

• Send every result for one turn in a single user message. Splitting them across several messages teaches the model to stop making parallel calls.
• Don't drop failed calls; return them with is_error: true.
• Order doesn't matter: results are matched by tool_use_id.

When to turn it off

If tools have side effects that must happen in order, set disable_parallel_tool_use: true in tool_choice, so the model makes at most one call per turn:

tool_choice: { type: 'auto', disable_parallel_tool_use: true }