Google is preparing to launch its new AI model, Gemini 3.8 Flash, as early as Wednesday, according to a report from The Wall Street Journal.
The launch is part of Google's effort to narrow the coding capability gap where it has lagged behind competitors OpenAI and Anthropic. The model, internally known as "Skimaki", reflects a concentrated push to strengthen coding performance in the increasingly competitive enterprise market.
In head-to-head internal testing on Google's Jetski developer platform, engineers reportedly preferred the upcoming model over Anthropic's flagship Claude Opus engine. "Internal tests of Gemini 3.8 Flash demonstrate narrowed coding-performance gap versus Anthropic and OpenAI," a person briefed on the results said.
While Gemini 3.8 Flash is not expected to match the scale of Google's flagship frontier models, its smaller architecture offers distinct operational advantages for rapid deployment. The lighter footprint requires significantly less compute capacity to modify, enabling multiple research teams to iterate simultaneously using reinforcement learning.
The launch comes less than three weeks after Gemini 3.7 Flash debuted on August 13, marking the third Flash update in under two months—a strategic shift toward targeted micro-updates rather than waiting months for full retrains.