Repository navigation
Price Vertex grounding and Sonar request fees - #97
Merged
Merged
Conversation
Two search fees were missing from the retrieval catalog. Grounding with Google Search on Vertex AI had no row; only the Gemini API channel carried the $14 per 1,000 queries price. Gateway-billed Vertex traffic shows every grounded call costing a whole multiple of $0.014 above its tokens, and Google's Gemini API pricing states the same per-query rate for Gemini 3.x. The Vertex row is dated from the earliest traffic it was checked against. Perplexity Sonar carried only `uncaptured_fees: true`. Perplexity charges a request fee on top of tokens by search context size: $5, $8 and $12 per 1,000 requests for low, medium and high, with low as the default. Gateway billing shows exactly $0.005 per request on traffic that names no size. The sonar token row keeps its flag until callers price requests alongside tokens. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The Vertex pricing page states the per-query price directly: Gemini 3 models are charged $14 per 1,000 grounding queries after 5,000 free a month, with billing from January 5, 2026. The row now carries that date and the page that states it. Gemini 2.5 and 2.0 stay per grounded prompt and are left out. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
VasiliyRad
marked this pull request as ready for review
October 5, 2026 22:59
The catalog guide says a catalog change updates the declared version, and the release gate ties the verified date to it. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
4 of 5 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Two search fees added to the
retrievalsection ofprices.yaml:google-vertex-ai/google_search_grounding: $14 per 1,000 grounding queries, effective 2026-01-05, Google's stated billing start for per-query grounding on Gemini 3 models.perplexity-api/search_request_low|medium|high: $5, $8 and $12 per 1,000 requests, effective 2025-04-18. Low is Perplexity's default when a request names no search context size.The
perplexity/sonartoken row keepsuncaptured_fees: true: nothing prices requests alongside tokens yet, so the row should still come outpartial.Evidence
cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing, read 2026-10-05): "Includes 5,000 Grounding Queries per month at no charge, aggregated across all Gemini 3 models. Grounding Queries exceeding those limits are billed at $14 per 1,000 Grounding Queries. You will be charged for each individual Grounding Query performed. Billing will start January 5, 2026." Gemini 2.5 and 2.0 are billed per grounded prompt at $35 per 1,000 after a daily allowance; they are not represented, since the retrieval catalog is keyed by channel and operation, not by model.google_searchtool show no such difference.The free monthly allowance on Vertex is not modelled; a customer's invoice can be up to $70 a month below what these rows compute.
Checks
pytest core/tests -q: 795 passed (Python 3.12) before the date change; retrieval and loader tests rerun after it.python core/tests/package/verify_artifacts.py: OK.versionandpricing_verified_atunchanged; only these rows were verified.Not in this PR
Applying retrieval prices to captured calls at ingest. The app prices tokens only, so these rows change nothing for customers until that lands.
🤖 Generated with Claude Code