Google's Gemini 3.5 Flash follows Anthropic and OpenAI in making newer AI models significantly pricier

The more pointed detail here is that Gemini 3.5 Flash, a model positioned in the efficiency tier, now costs more on agent workloads than the previous flagship Pro model. That inversion suggests the pricing pressure isn't just about raw capability jumps but about how multi-step agentic usage patterns are becoming the new billing surface labs are optimizing around.

This is largely disconnected from recent activity in our archive, as we have no prior coverage to anchor it to. That said, it belongs to a broader pattern worth naming: all three major labs have now moved in the same direction on inference pricing within a compressed timeframe, which is less likely to be coincidence than a shared read on enterprise willingness to pay. When Anthropic and OpenAI raised prices on their respective newer models, the framing was capability justification. Google is now using the same framing, which normalizes the trajectory rather than challenging it.

Watch whether enterprise API consumption data (from cloud resellers or earnings calls in Q3 2026) shows volume compression at the new price points. If usage growth flattens despite capability gains, the labs will face a real test of whether this pricing holds or triggers a race back down.