Gemini 3 Flash is priced at $0.50 per 1M input tokens and $3 per 1M output tokens, compared to Gemini 2.5 Flash's $0.30 per 1M input and $2.5 per 1M output
Gemini 3 Flash is our latest model with frontier intelligence built for speed that helps everyone learn, build, and plan anything — faster.
Context & Ripple Effects
Gemini’s Flash line subsequently split into a lower-cost Lite option that Google said outperformed 2.5 Flash, while later Flash releases moved pricing up and then down again. The initial premium over 2.5 Flash is therefore an early marker of a more explicit performance-and-cost ladder, including the Gemini 3.1 Flash-Lite launch and lower 3.6 Flash pricing versus 3.5 Flash.
First-order effects
- Gemini API customers moving from 2.5 Flash to Gemini 3 Flash face higher listed token costs: input pricing rises from $0.30 to $0.50 per million tokens and output pricing from $2.50 to $3.
- The increase makes workload mix more consequential immediately: applications that generate substantial output see a smaller proportional increase than input-heavy ones.
Second-order effects
- Developers gain a clearer incentive to segment workloads by capability and cost rather than treating the Flash family as a single default tier; the later Flash-Lite offering reinforces that choice.
- Google’s pricing creates room for a cheaper sibling model to serve price-sensitive use cases while reserving the newer Flash tier for customers that value its stated speed and intelligence.
Third-order effects
- If this pattern holds, lightweight model families will be sold as multi-tier portfolios, with model selection increasingly based on effective cost for a useful task rather than a single headline token rate.
- Frequent repricing across adjacent model versions may make token pricing less durable as a buying signal, pushing customers toward routing, benchmarking, and workload-specific cost controls.
The trend: Frontier-model providers are turning “fast” model lines into tiered price-performance portfolios rather than maintaining one broadly priced low-cost default.