Google boosts reasoning in Gemini 3.8 Flash without raising the price

Google has released a new model: Gemini 3.8 Flash. It is the third Flash model in just six weeks, so the pace at Google is high right now.

What is new is better reasoning and coding without costing more. The price is $0.75 per million input tokens and $3.75 per million output tokens, the same level as its predecessor.

There is also a special edition called Gemini 3.8 Flash Cyber, aimed at cybersecurity. It finds vulnerabilities and fixes them automatically and is available through Google's Fairwind program.

On the broad benchmarks, Anthropic's Claude Opus 5 is still ahead. But in specific domains such as law, finance, biology and coding in terminal environments, Gemini 3.8 Flash beats both Opus 5 and OpenAI's GPT-5.6 Sol.

And here is the detail that actually matters for your wallet: the competitors cost 4 to 5 times as much per token. Opus 5 is at $5, GPT-5.6 Sol at $4, versus Gemini's $0.75.

What does it mean for you? If you or your company use AI for coding, legal or financial research, this is worth looking at now, especially if you pay for API access and want to keep costs down.

If you just use a chatbot day to day, you can safely wait. It will not be visible to you right away, but the price pressure may eventually mean cheaper AI subscriptions in general.

What it means for you

If you work with coding, law or finance and pay for API access yourself, this could be an obvious opportunity to save money now. Check whether your current setup can switch to Gemini 3.8 Flash without hassle. If you just use a chatbot day to day, you will not notice a difference right now, but lower prices at Google may over time pull the rest of the market down.