The Model Beating OpenAI on Search-Grounded Pricing Is Cheaper Than You Think

Based on the public pricing sheets checked on March 15, 2026 in our broader AI token pricing comparison, the short answer is straightforward: xAI currently has the cheapest clearly exposed first-party search fee in the comparison.

That does not make this the universal best buy. It makes it the cleanest answer to one narrow question: which provider has the clearest advantage once your workflow calls first-party search on every turn. That distinction matters because a lot of teams still confuse the cheapest model row with the cheapest production stack.

The short answer

OpenAI and Anthropic both sit at $10 per 1,000 search calls. Google Gemini 3 sits at $14, and Gemini 2.5 grounding is $35 per 1,000 grounded prompts. xAI Web Search and X Search both sit at $5 per 1,000 calls.

That fee can matter more than people expect because search-heavy products trigger it over and over again. Once every turn hits search, the “cheap model” story can get flipped by the tool line very quickly.

The pricing rows that matter

Provider	Search fee	Notes
xAI	$5 per 1,000 calls	Cheapest clearly exposed first-party search row.
OpenAI	$10 per 1,000 calls	Search content tokens can also apply.
Anthropic	$10 per 1,000 calls	Normal token charges still apply.
Google Gemini 2.5	$35 per 1,000 grounded prompts	Meaningfully more expensive.

On a search-heavy assistant, that tool line is not decorative. It becomes part of the core bill. That is why xAI can look much more competitive in grounded-answer products than in model-only comparisons.

Why the headline can mislead

Cheap search does not automatically mean cheapest whole product. The model layer, output volume, and attachment or collection state can still move the ranking.

It also does not settle quality. The pricing win is real, but the workflow still has to be good enough for your use case. Cheap grounded answers that miss the user’s real job are still expensive.

When this is the right pick

• your product uses search on most turns
• you want predictable search fees before adding token costs
• you are testing grounded-answer economics rather than buying a deep managed stack

When to ignore the headline

• your workflow rarely uses search
• you need Google’s broader ecosystem more than the cheapest search line
• your app becomes provider-specific through collections and other native state

Bottom line

For search-heavy products, the biggest pricing surprise is that xAI’s tool fee is meaningfully lower than the more visible alternatives. That is enough to change who deserves a serious trial.

If you want the wider market context, start with the full provider-by-provider pricing breakdown and, for media-specific workloads, the separate image and video generation API comparison.