Google introduced Gemini 3 Flash on Wednesday, appealing to enterprises that want to use Gemini 3 without the high cost. The new model shows that Google is capitalizing on its current popularity among enterprise customers, but it’s also another example of how model makers like Google are offering cheaper models that are on par with their Frontier models.
The Gemini 3 Flash joins a highly anticipated family of models that was introduced last month with the Gemini 3 Pro and Gemini 3 DeepThink. Flash uses similar logic gemini 3 proBut uses fewer tokens to complete everyday tasks. Google said it can also adapt its thinking based on the use case it is being used for.
And it comes at an affordable price. For paying users, Gemini 3 Flash is priced at $0.50 per million tokens input for text, image, and video, and $1 per million tokens for audio inputs. The output is $3 per million tokens. By comparison, Gemini 3 Pro inputs are between $2 and $4 per million tokens. Output prices range from $12 to $18 per million tokens.
model, which will replace gemini 2.5 flash The cloud provider said Gemini apps achieve the same coding level as Pro with lower latency. Like other models in the Gemini 3 family, the flash tool is suitable for use and has multimodal capabilities. Some use cases for models include video analysis and data extraction.
“Since its launch, Gemini 3 has become much more attractive to developers who are looking for multimodal experiences,” said Lian Ze Su, an analyst at Omdia, a division of Informa TechTarget. “We’re seeing Google become more capable of delivering cutting-edge multimodal AI experiences.”
According to analysts, the Gemini 3 demonstrates Google’s efforts to meet the needs of multiple enterprises, considering factors such as flash accuracy, response quality, cost, and speed when determining which model best suits their use case.
While some use cases may require the use of gemini pro modelThe majority can rely on the flash model, said Gartner analyst Arun Chandrasekaran.
“It’s not like you’re getting an inferior model at a very low price,” Chandrasekaran said. “It may be that there are some hard logic-oriented use cases where you would use the Pro model, but for many other tasks, flash will provide a perfect balance between performance versus speed and cost.”
He said this model is in line with a broader strategy among providers, which is to “take a lot of details away from the user”.
In other words, model makers want to get to a point where users can’t determine which model will answer their query unless they make an explicit selection, Chandrasekaran said.
“It’s also a way for (model makers) to reduce their costs,” he said. “If they can generate more responses from a lower-cost model, of course they will do that.”
While offering a lower-cost model for enterprises provides an option, Google will face a challenge in getting users to choose between Pro and Flash.
“The way you present it would be better if there were excellent talking points and swim lanes,” Chandrasekaran said. “Given the insignificant differences at some levels between these models, this will always remain a challenge.”
