Google’s next big AI model is slipping
Google helped ignite the modern AI race, but staying on top has proven far harder than joining it. A new Bloomberg report claims the company has fallen months behind its internal schedule for launching Gemini 3.5 Pro, its next flagship model. The culprit? Coding.
Engineers are reportedly still wrestling with the model’s ability to write and debug software — a weakness that has become impossible to ignore. The delay isn’t just about polishing a chatbot. It exposes a deeper problem: Google’s sprawling structure, competing divisions, and stricter safety reviews are slowing its response to rivals who move at breakneck speed.
While OpenAI, Anthropic, and Meta keep shipping increasingly capable models, Google appears stuck balancing innovation with the trust it has built across products used by billions. That balancing act is getting harder by the week.
Coding remains Gemini’s biggest weakness
Bloomberg’s report, citing multiple current and former Google employees, says Gemini 3.5 Pro has been delayed because the company hasn’t hit its internal targets for coding performance. Google even refreshed the model’s training data late last month to boost its coding chops. The results? Reportedly underwhelming.
This matters because writing code has become the clearest benchmark separating today’s leading AI models. OpenAI, Anthropic, and Meta have all poured resources into developer-focused systems that can write, debug, and reason through complex software projects. According to the report, both OpenAI and Meta currently outperform Google’s available models in this arena.
Google, for its part, insists progress is happening. In a statement cited by Bloomberg, the company said it is testing Gemini 3.5 Pro, an upgraded Flash model, and other AI systems with partners. It also noted ongoing discussions with the US government around testing standards and AI safety.
The timing is awkward. Many observers expected Gemini 3.5 Pro to debut at Google I/O earlier this year. Instead, Google offered incremental Gemini updates while competitors kept shipping frontier models.
Google’s scale is a double-edged sword
Unlike most AI startups, Google isn’t building models in isolation. Every major Gemini release eventually needs to work across Search, YouTube, Maps, Android, Workspace, Cloud, and dozens of other products. That scale gives Google massive advantages — including unmatched access to real-world data. But it also creates layers of internal coordination that can slow decision-making to a crawl.
Employees describe competing priorities across DeepMind, Google Cloud, Android, and other teams. Overlapping AI coding efforts make it harder to maintain a unified strategy. Former employees also mentioned internal disagreements over AI-generated code, plus earlier restrictions on using Gemini for software development, which limited experimentation during the tech’s early rollout.
Policy shifts and internal friction
Google says those policies have evolved. The company claims roughly 75 percent of its production code is now generated using AI. Internal coding tools are being consolidated under a common platform called Google Antigravity. Engineers are now expected to use AI for coding, although some still face computing capacity constraints due to intense internal demand for GPU resources.
But frustration is simmering. Some researchers have reportedly left for competitors like Anthropic. Customers are split on Gemini 3.5 Flash, too. Companies like Figma praise its balance of speed and quality. Others, including education platform Platzi, say it sits in an awkward middle ground — higher costs than previous Flash models without matching the reasoning power of premium rivals.
The real question for Google
The bigger picture is that Google’s AI challenge is no longer about proving it can build frontier models. Few doubt that it can. The real question is whether a company of Google’s size can ship those models quickly enough in an industry where competitors now measure progress in weeks, not months.
For now, the Gemini 3.5 Pro delay is a telling sign. Google’s next move will reveal whether it can adapt its internal machinery to the pace of the AI race — or whether it gets left behind in the very revolution it started.
Want to stay updated on Google’s AI moves and the broader AI model competition? Keep an eye on this space. Also check out our breakdown of Google Antigravity coding tools and how they fit into the bigger picture of developer-focused AI systems.
