Google Releases Gemini 3.8 Flash — Third Model in Six Weeks — Holds Introductory Pricing
The short answer
Google's latest Flash model adds more reasoning steps and iterative tool use at the same opening price of $0.75 per million input tokens, but the company signals the rate may not hold. The iterative tool-calling capability is the most practically significant change for businesses using AI to automate multi-step work — tasks like pulling data from multiple sources, running calculations, or completing research that requires several lookups in sequence.
- Google Gemini
- AI pricing
- API
- model release
- workflow automation
What happened
Google has released Gemini 3.8 Flash, the third iteration of its Flash model line in approximately six weeks. The company says the new model performs more reasoning steps on complex tasks than its predecessor and calls tools iteratively — meaning it can take multiple sequential actions to complete a task rather than responding in a single pass. Google described this as the model 'working harder' on difficult inputs.
Introductory pricing matches the previous version: $0.75 per million input tokens and $3.75 per million output tokens. Google has indicated final pricing may be higher given the model's greater computational demands, but has not announced when introductory rates will expire. Updates to Google's Pro model line have been less frequent during this same period.
What it means for your business
The iterative tool-calling capability is the most practically significant change for businesses using AI to automate multi-step work — tasks like pulling data from multiple sources, running calculations, or completing research that requires several lookups in sequence. At the same introductory price as the prior version, testing the new model against your current workflows carries no additional short-term cost if you are already using Gemini through Workspace or the API. Pay attention to Google's signal that pricing may rise: if you are building automated workflows on the API, document your current token usage now so you can evaluate the cost impact when rates change. The rapid release cadence — three models in six weeks — also suggests that designing workflows around stable task structures, rather than a specific model version, is the more durable approach.
Source: The Verge