Web search tool pricing in 2026

One call for every model

A built-in search tool runs inside its vendor's API. The SpecStream Search API is a plain REST call, so the same search, price and reply serve GPT, Claude, Gemini, Grok or an open model, and changing models leaves your search code as it is.

curl -G https://api.specstream.ai/v1/web-search \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" \
  --data-urlencode "q=Who is the CEO of Mistral AI?"

const reply = await fetch(
  "https://api.specstream.ai/v1/web-search?" + new URLSearchParams({ q: "Who is the CEO of Mistral AI?" }),
  { headers: { Authorization: `Bearer ${process.env.SPECSTREAM_API_KEY}` } },
).then((r) => r.json());

const { answer, sources, price_usd } = reply.data;

Return data.answer and data.sources to your model as the tool result. Agents that speak MCP add https://api.specstream.ai/mcp and call the search_web tool.

What 100,000 searches cost

100,000 searches cost $500 on SpecStream, with no tokens. The same volume costs $500 plus tokens on xAI, $1,000 plus tokens on OpenAI and on Claude, $1,330 on Gemini 3 grounding after the 5,000 free a month, and $1,400 on Grounding with Bing. Each SpecStream reply carries price_usd, and the account's request log and usage totals show every call and its price.

Results your code reads

The built-in tools hand their results to the model. Grounding with Bing returns none of its tool output to developers, and Claude's page content comes back encrypted. A SpecStream search returns JSON your code reads without a model in between: answer with the page it was read on and its read time, rows with the facts the request asks for, and sources in rank order, each marked cited when a value comes from it. Add passages=true and each source carries up to about 3,500 characters of its page.

The most accurate web search API in the world

The SpecStream Search API scores a 97.9 F-score on all 1,000 official SimpleQA Verified questions, with 978 correct. On WideSearch it scores 64.0 item F1 and 43.4 row F1 on all 200 official tasks. The scores and the method are at specstream.ai/benchmarks.

Busy calls cost nothing

When search is busy, /v1/web-search answers 503 with a Retry-After header and charges nothing, and the Python and TypeScript SDKs send it again once by themselves when the wait is 10 seconds or less. Your code decides how many searches run, and each one costs $0.005.

Search, pages and rows on one key

The same balance pays for the Extract API at $0.80 per 1,000 pages and for Deep Search, which takes a request in plain words and returns every match on the live web as structured rows at 5¢, 15¢ or $1 a row, under a maximum you approve before it runs. Search APIs priced by the call are on web search API pricing.

SpecStream

The live web data engine for teams and agents. The Search API costs $5 per 1,000 searches, bills no tokens, works with any model and scores 97.9 F-score on all 1,000 official SimpleQA Verified questions. The Extract API costs $0.80 per 1,000 pages. Deep Search returns every match as structured rows at 5¢, 15¢ or $1 a row.

OpenAI web search

$10 per 1,000 calls plus search content tokens at the model's rates, or $25 per 1,000 for web search preview on other models, inside an OpenAI response. SpecStream gives OpenAI models the same search at $5 per 1,000 with no tokens. SpecStream as an OpenAI web search alternative.

Claude web search tool

$10 per 1,000 searches on the Claude API, plus standard token costs for the content it brings back. SpecStream gives Claude the results at $5 per 1,000 as a client tool or through Anthropic's MCP connector. SpecStream as a Claude web search tool alternative.

Gemini grounding with Google Search

$14 per 1,000 search requests after 5,000 free a month on Gemini 3 and newer, and $35 per 1,000 grounded prompts after 1,500 free a day on Gemini 2.5. SpecStream returns the ranked results to your code at $5 per 1,000. SpecStream as a Grounding with Google Search alternative.

Grounding with Bing Search

$14 per 1,000 transactions in Azure AI Foundry Agent Service, up to 150 a second and 1 million a day, and the tool output stays with the agent. SpecStream returns the results to your code at $5 per 1,000. SpecStream as a Grounding with Bing Search alternative.

xAI web search

$5 per 1,000 calls of web_search on the xAI API, with token usage billed on top and Grok deciding how many searches run. SpecStream costs the same $5 per 1,000 with no tokens, and returns the results to Grok or any other model. SpecStream as an xAI web search tool alternative.