Pricing
USD per 1,000 searches, answers or pages fetched.
| Service | Search | Ask / answer | Fetch |
|---|---|---|---|
| Scrapegoat | $1.00 | $1.60 | Not published |
| Parallel ↗ | $1.00 | $50.00 | $1.00 |
| Exa ↗ | $7.00 | $5.00 | $1.00 |
| Firecrawl ↗ | $1.498 | — | $0.749 |
| Gemini 3+ ↗ | $14.00* | Variable* | — |
| Perplexity ↗ | $5.00 | Variable† | $0.50† |
*Gemini Search Grounding: per 1,000 model-issued searches, plus model tokens, after 5,000 free searches/month. One answer may use multiple searches. Billing details ↗
†Perplexity answers use model tokens plus tool fees. Fetch is an Agent API tool fee per 1,000 invocations; model tokens are additional.
Plans, sources and comparison details
Selected modes: Parallel Fast Search, Responses Medium and Extract; Exa Developer with Auto Search, Answer and Contents; Firecrawl Scale at $749/month for 1M credits.
Comparison uses the largest publicly priced subscription where available: Firecrawl Scale, billed monthly. Exa and Parallel publish pay-as-you-go rates rather than priced volume tiers; the selected Search and answer modes are identified above. Negotiated enterprise rates are not public and are excluded.
Search assumes up to 10 results. Firecrawl uses 2 credits per search and 1 per basic page scrape; rates assume the full monthly allowance is used. Annual Scale pricing is $599/month, equivalent to $1.198 per 1,000 searches or $0.599 per 1,000 pages.
Parallel Responses ranges from $10 (Low) to $250 (High) per 1,000 answers; Medium is $50. Parallel Advanced Search is $5; Exa Lite Search is $4 and Deep Search is $12 per 1,000 searches. These modes differ in research depth and latency. Exa Contents is $1 per 1,000 pages per content type. Firecrawl has no directly equivalent cited-answer endpoint listed here.
Gemini uses the Developer API paid tier; no public volume discount is assumed. Its 5,000-search monthly allowance is shared across Gemini 3 and newer models. After that allowance, 1,000 answers averaging one search each incur $14 in grounding fees; three searches each incur $42. Model input and output token charges are additional. Google Search Grounding is a model tool, not a standalone ranked-results or page-fetch API. Gemini 2.5 pricing instead lists $35 per 1,000 grounded prompts after its applicable free allowance.
Perplexity uses published pay-as-you-go rates: standard Search API is $5 per 1,000 successful requests; Fast Search is $1. Each request can contain up to five queries, with no additional token charge for standalone Search. Agent API answers are usage-based: standard web search costs $2.50 per 1,000 tool invocations (Fast Search $1), and fetch_url costs $0.50 per 1,000 invocations, plus the selected model’s token charges. The fetch price is not an all-in standalone page retrieval rate. Custom enterprise rates are excluded.
Capabilities, freshness and output differ. Search content, extra results, AI extraction, summaries and other add-ons may cost more. Prices exclude taxes. Scrapegoat cache hits count as requests.