Alphabet shares sank roughly 4% on Thursday after a report that Google has once again delayed releasing its flagship Gemini 3.5 Pro artificial intelligence model, with the model's coding capabilities falling short of internal expectations even as rivals accelerate.
The Gemini 3.5 Pro model is now months behind schedule, according to Bloomberg, which cited sources familiar with the matter. For readers tracking the latest AI developments, the setback underscores how the global AI model race is intensifying — and AI Buzz Wire continues to cover every major model launch and delay as it happens.
Google first announced Gemini 3.5 Pro at its annual Google I/O developer conference in May 2026, saying at the time that the model was being used internally but would not be ready for a broader rollout until the following month. That target has now slipped repeatedly, and the project has dragged on well beyond its original summer window.
Coding Performance Falls Short
According to the Bloomberg report, Gemini 3.5 Pro's coding capabilities in particular were short of what Google's own engineers expected. Code generation has become one of the most important and lucrative use cases for AI model providers, and the company's internal benchmarks reportedly showed the model underperforming just as competitors pushed ahead with their own releases.
The timing is uncomfortable for Google. Rivals like OpenAI and Meta have recently debuted new AI models that outpace Google's current offerings in generating software code.
Meta last week released its Muse Spark 1.1 model, which the company's AI chief Alexandr Wang described as the social media giant's "strongest model for agentic and coding work yet." OpenAI, meanwhile, launched its GPT-5.6 Sol model, with CEO Sam Altman saying it is 54% more token efficient on agentic coding tasks — a direct pitch to developers who increasingly judge models on cost relative to performance.
Chinese labs are also crowding the field. Companies like Z.ai now offer open-weight variants that developers can access for free through the open-source ecosystem, adding further pressure on Google to ship a competitive coding model.
Alphabet Responds
An Alphabet spokesperson pushed back against the notion that the company is falling behind, telling CNBC in an emailed statement that the company is "shipping quickly across a wide range of models while keeping them highly cost-effective for customers."
"We're currently testing 3.5 Pro, an upgraded Flash model, and other models with partners, and we're productively engaged with the U.S. government," the spokesperson said, an apparent reference to ongoing discussions between frontier AI developers and regulators in Washington.
Why the Coding Race Matters
The intense focus on coding performance reflects a broader shift in how AI models are being evaluated and sold. While early chatbot benchmarks measured general knowledge and reasoning, the attention of enterprise buyers and developers has increasingly moved to agentic coding — the ability of a model to write, debug, and execute software across multiple steps.
For Google, the stakes are unusually high. The company's Gemini family is meant to anchor its consumer products, its cloud offerings, and its developer tools all at once. A model that lags on coding risks weakening Google Cloud's pitch to enterprise customers who are weighing it against alternatives from Anthropic, Microsoft-backed OpenAI, and a growing field of open-weight competitors.
Analysts have warned that repeated delays can create a perception problem even when the underlying technology is competitive. Google's earlier Gemini 3.5 Flash release added built-in computer-use capabilities for cross-platform AI agents, signaling that the company is still shipping in the Gemini 3.5 family — just not yet the flagship Pro tier that would anchor its most premium offerings.
Market Reaction
Thursday's sell-off erased billions in market value from Alphabet, reflecting investor sensitivity to any sign that Google's multi-billion-dollar AI investment is not translating into timely product releases. The decline came on a day when broader tech markets were also under pressure, amplifying the move.
The delay also complicates Google's competitive positioning heading into the second half of 2026, when several major labs are expected to release their next-generation models. With OpenAI, Meta, and Chinese labs all pushing forward, the window for Google to reassert leadership in coding and agentic tasks is narrowing.
For now, the company says it will continue testing Gemini 3.5 Pro with partners. But with no firm release date, the question hanging over Mountain View is simple: when the model finally does ship, will it still be good enough to lead?
Stay Ahead of the AI Model Race
The competition between Google, OpenAI, Meta, and a fast-growing field of Chinese labs is reshaping the entire technology landscape. Don't miss a beat — read more breaking AI news and in-depth analysis at AI Buzz Wire.
Read more AI news →
