Google Flex mode, 50% off standard API rates

Google Flex mode is an inference tier (preview) that bills per token at 50% of standard API rates when service_tier is set to flex in a paid Gemini API tier.In exchange for lower request-processing priority, reliability is lower and response latency is longer.Google Cloud (Gemini Enterprise Agent Platform) also offers Flex PayGo (Pre-GA), which is 50% cheaper than Standard PayGo.

What you get

Billed per token at 50% of standard API rates in the Flex inference tier (preview) of the Gemini API paid tier

Price

In a paid Gemini API tier, setting service_tier to flex bills tokens at 50% of standard API rates. Google Cloud (Gemini Enterprise Agent Platform) Flex PayGo is also 50% cheaper than Standard PayGo. No separate coupon code is required.

Valid until

No official deadline has been announced. The Gemini API Flex inference tier was announced on April 2, 2026, and is in preview. Google Cloud Flex PayGo is in the Pre-GA stage.

Requirements

- It is available in all Gemini API paid tiers, with no separate coupon code.- In Gemini API, service_tier must be set to flex in the request for the 50% standard API rate to apply.If it is not specified, the request is handled at the standard tier.- In exchange for lower request-processing priority, reliability is lower and response latency is longer.- Gemini 3.8 Flash paid-tier rates (USD per 1 million tokens) through December 31, 2026 are Input, Standard $0.75, Flex $0.375; Output (including thinking tokens), Standard $3.75, Flex $1.875.From January 1, 2027, the input rate increases to Standard $1.50, Flex $0.75.- The Gemini API Flex inference tier was announced on April 2, 2026, and is in preview.As of the documentation update on November 12, 2026 (UTC), Google Cloud (Gemini Enterprise Agent Platform) Flex PayGo is Pre-GA, a stage where support may be limited.- In Gemini API, it can be used with GenerateContent and Interactions API requests.Supported models include Gemini 4 Carbon and Gemini 3.8 Flash.- Google Cloud (Gemini Enterprise Agent Platform) Flex PayGo supports only Gemini 3.8 Flash and earlier models.- Gemini API documentation does not mention regional restrictions.Google Cloud Flex PayGo does not support regional or multi-region endpoints.

Check the conditions on the official page ↗