Google Cloud (Vertex AI ranking API)
Status GA is inferred: the docs list every model as available with discontinuation "To be determined" and do not label any as preview. Records that exceed the per-record token limit are truncated, not rejected.
Google Cloud rerankers - verified
Google Cloud (Vertex AI ranking API): 6 reranker models, read from the provider's own pages on October 18, 2026. Google Cloud rerankers are priced per 1,000 queries, from $1.00 per 1,000 queries for all 6 models.
Read from cloud.google.com. Compare every reranker
Status GA is inferred: the docs list every model as available with discontinuation "To be determined" and do not label any as preview. Records that exceed the per-record token limit are truncated, not rejected.
semantic-ranker-default-005 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-default-005 · Yes · No · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
1,024 tokens
per record (title plus content)
“semantic-ranker-default-005 · Yes · No · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
25 languages
“semantic-ranker-default-005 · Yes · No · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
Current target of the semantic-ranker-default@latest alias. Released September 1, 2026.
semantic-ranker-fast-005 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-fast-005 · No · Yes · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
1,024 tokens
per record (title plus content)
“semantic-ranker-fast-005 · No · Yes · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
25 languages
“semantic-ranker-fast-005 · No · Yes · Text (25 languages) · 1024 · September 1, 2026 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
Current target of the semantic-ranker-fast@latest alias. Released September 1, 2026.
semantic-ranker-default-004 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-default-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
1,024 tokens
per record (title plus content)
“semantic-ranker-default-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
25 languages
“semantic-ranker-default-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
Superseded as the default alias by 005; no discontinuation date set. Released April 9, 2025.
semantic-ranker-fast-004 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-fast-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
1,024 tokens
per record (title plus content)
“semantic-ranker-fast-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
25 languages
“semantic-ranker-fast-004 · No · No · Text (25 languages) · 1024 · April 9, 2025 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
Superseded as the fast alias by 005; no discontinuation date set. Released April 9, 2025.
semantic-ranker-default-003 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-default-003 · No · No · Text (25 languages) · 512 · September 10, 2024 · To be determined”
512 tokens
per record (title plus content)
“semantic-ranker-default-003 · No · No · Text (25 languages) · 512 · September 10, 2024 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
25 languages
“semantic-ranker-default-003 · No · No · Text (25 languages) · 512 · September 10, 2024 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
The prose on the same page calls the 003 model semantic-ranker-512-003 while the table lists semantic-ranker-default-003. Released September 10, 2024.
semantic-ranker-default-002 - API - Per 1,000 queries
$1.00 per 1,000 queries
“Ranking $1.00 / 1,000 count”
Note: Pay-as-you-go column; the next two columns ($0.90 and $0.80) are 1-year and 3-year savings-plan rates. A query is up to 100 documents; every further 100 documents counts as another query.
GAinferred
“semantic-ranker-default-002 · No · No · Text (en only) · 512 · June 3, 2024 · To be determined”
512 tokens
per record (title plus content)
“semantic-ranker-default-002 · No · No · Text (en only) · 512 · June 3, 2024 · To be determined”
1,000
hard limit
“You can include up to 1000 records per request.”
English only
“semantic-ranker-default-002 · No · No · Text (en only) · 512 · June 3, 2024 · To be determined”
No (API only) inferred
Note: No official open weights are listed in the sources checked; the provider documents this as an API model.
English only. Released June 3, 2024.
semantic-ranker-default-005: $1.00 per 1,000 queries. semantic-ranker-fast-005: $1.00 per 1,000 queries. semantic-ranker-default-004: $1.00 per 1,000 queries. semantic-ranker-fast-004: $1.00 per 1,000 queries. semantic-ranker-default-003: $1.00 per 1,000 queries. semantic-ranker-default-002: $1.00 per 1,000 queries. Prices are in Google Cloud's own units as of October 18, 2026.
semantic-ranker-default-005: 1,024 tokens, per record (title plus content). semantic-ranker-fast-005: 1,024 tokens, per record (title plus content). semantic-ranker-default-004: 1,024 tokens, per record (title plus content). semantic-ranker-fast-004: 1,024 tokens, per record (title plus content). semantic-ranker-default-003: 512 tokens, per record (title plus content). semantic-ranker-default-002: 512 tokens, per record (title plus content).
No official open weights were found in the sources checked; Google Cloud documents these as hosted models.
Last verified from each provider's own docs, pricing pages and Hugging Face cards; every quote was string-matched against the page that day. Prices are shown in the provider's own unit and change without notice, so check the linked source before you commit a budget.
Newsletter
One email per week. New model releases, retrieval and pricing updates and what they mean for what you build.