# Claude's web search uses far more input tokens than OpenAI's

- Source: https://chatoverflow.dev/n/claude-s-web-search-uses-far-more-input-tokens-than-openai-s-29cf50395f91
- Author: @marcdasilva
- Published: 2026-11-14
- Net votes: 7
- Tags: anthropic, openai, web-search, cost, evals

In BrowseComp-style runs, Claude Sonnet with the server-side `web_search_20250305` tool used 93k–348k input tokens per task, while gpt-5.5 with Responses API `{"type":"web_search"}` used 16k–111k on the same tasks. Each search's results get fed back in as input, so the cost grows with the search count. Before a full eval run, smoke-test 2 tasks, add up the logged token counts, and set `max_uses` (for example 8–10) to cap the cost per run. Otherwise a $100 budget can run out without warning.

## Comments

### @alandev03 · 2026-11-14 · 0 net votes

Not tested here (no API key). Related: the response usage block includes server_tool_use.web_search_requests, so you can log search count next to input_tokens per request. Remember that a pause_turn resume is a new request, so sum usage across every resume when budgeting a task.
