Google's Gemini 3.5 Pro is months behind schedule as coding performance lags rivals, Bloomberg reports

Google's next flagship model, Gemini 3.5 Pro, is running months behind its internal schedule, according to Bloomberg, which cited people familiar with the matter. The company has been holding the release back to keep improving the model's capabilities, particularly in coding.

What's new

Bloomberg reports that "Alphabet Inc.'s Google is months behind schedule on delivering Gemini 3.5 Pro, its most powerful flagship AI model, because the company has been taking time to try to improve its capabilities, particularly in coding, according to people familiar with the matter."

The report, by Bloomberg's Julia Love and Davey Alba, says the slippage has caused internal friction: "The delay has been a source of frustration for Google engineers, AI researchers and managers, many of whom are concerned the company risks losing an edge in the market as rivals Anthropic and OpenAI produce models that exceed Gemini's capabilities, according to 10 current and former employees." Bloomberg attributes part of the delay to Google's release process itself, noting the company "has multiple layers of stakeholders involved in preparing models for release, working to weave AI across a vast product portfolio, including search, maps and YouTube, which can cause delays."

Context

Google has been shipping steadily on its current Gemini 3.x line — Gemini 3.5 Flash is generally available today, alongside Gemini 3.1 Pro Preview — while coding has become the most closely watched battleground among frontier labs. Anthropic and OpenAI have both leaned hard into agentic coding this year, from Claude's coding-focused releases to OpenAI's GPT-5.6 family and its Codex tooling, and both are frequently cited as the models Gemini 3.5 Pro needs to beat. A multi-month slip on a flagship model, attributed by Bloomberg's sources specifically to coding shortfalls, suggests Google's internal bar for Gemini 3.5 Pro was set against that competition rather than against its own prior-generation models.

Why it matters

If accurate, this is a rare acknowledgment — via ten current and former employees, not a company statement — that Google's product cadence is being constrained by its own engineers' assessment that the model isn't yet good enough versus Anthropic and OpenAI's latest releases, specifically on coding. That's a notable admission for a company that has otherwise been shipping frequently across its Gemini line this year. It also illustrates a structural tension Bloomberg's sources point to directly: Google's advantage of embedding AI across search, Maps, YouTube, and its broader product portfolio also means more stakeholders and more coordination before any single model can ship, which can slow exactly the kind of frontier release cycle competitors without that portfolio don't have to manage.