For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.Source for Data retention (opens in a new tab)
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$1.50
OUTPUT / 1M TOKENS$7.50
standard global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional. Cached input price: $0.15 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$1.65
OUTPUT / 1M TOKENS$8.25
standard us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional. Cached input price: $0.165 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$2.70
OUTPUT / 1M TOKENS$13.50
priority global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional. Cached input price: $0.27 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$2.97
OUTPUT / 1M TOKENS$14.85
priority us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional. Cached input price: $0.297 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$0.75
OUTPUT / 1M TOKENS$3.75
flex global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Flex rate from the combined Batch/Flex table; batch is explicitly unsupported. Explicit cache storage is additional. Cached input price: $0.075 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Exact default thinking level for this Cyber variant is not published in the reviewed card.
INPUT / 1M TOKENS$0.825
OUTPUT / 1M TOKENS$4.125
flex us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Flex rate from the combined Batch/Flex table; batch is explicitly unsupported. Explicit cache storage is additional. Cached input price: $0.0825 / 1M tokens (base tier).
For ordinary Google models under Cloud online terms, suspicious activity may trigger prompt logging for up to 90 days and human review, with no model training. Master Agreement customers are exempt from that prompt logging by default. Abuse logs are not CMEK encrypted; separate caching rules still apply.
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
standard global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
Google Enterprise Agent Platform (standard, us) · standard
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
standard us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
Google Enterprise Agent Platform (priority, global) · priority
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
priority global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
Google Enterprise Agent Platform (priority, us) · priority
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
priority us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
Google Enterprise Agent Platform (flex, global) · flex
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
flex global rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Flex rate from the combined Batch/Flex table; batch is explicitly unsupported. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
Google Enterprise Agent Platform (flex, us) · flex
Generic guide gives 4,096 for Gemini 3 but 6,144 for named Flash exceptions. Flash Cyber is not explicitly included in the exception list, so its exact minimum remains unverified.
Published Flash Cyber rates give a 90% cached-read discount. Explicit storage is $1 per million token-hours, separately from cached read charges.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Caches are project-isolated. Explicit blob/text content is limited to 10 MB. Avoid modifying referenced Cloud Storage objects until caches expire or are deleted. Provisioned Throughput caching is preview.Source for Limits & data boundaries (opens in a new tab)
flex us rates, output includes thinking. Cyber variants do not receive the general-model introductory promotion. Flex rate from the combined Batch/Flex table; batch is explicitly unsupported. Explicit cache storage is additional.
Converted exactly from $0.000001 per token-hour. Storage is separate from input, output and cached-read charges.
Observed 14 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 14 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
Google explicitly describes a post-training version of Gemini 3.8 Flash with a distinct proprietary hosted selector and no supplied open weights.
