Gemini users waited.
And waited.
Some might even say they really waited.
At Google I/O in May, the hope was real. Sundar Pichai, Google’s CEO, told a crowd of developers and press that Gemini 3.5 Pro would hit the streets the following month. It sounded concrete. He promised it was internal-tested, showing “great improvements,” and ready to roll out by June.
Here it is mid-July.
Silence.
Instead of the heavyweight champion everyone expected, Google launched Gemini 3.5 Flash. It’s lighter. It’s for everyday use. But it’s not the flagship model people were circling on their calendars.
Why is Google’s Gemini 3.5 Pro taking so long to release?
Bloomberg reporters Julia Love and Davey Alibaba broke down the messy reality behind the delay yesterday. The frustration inside Mountain View is palpable. Engineers, AI researchers, managers—everyone is worried. They think Google might be losing its edge. Not to a slow giant, but to rivals who move faster. Anthropic. OpenAI.
Mashable asked Google about the timeline. The answer? Same as Bloomberg got. A PR statement wrapped in corporate safety language.
“We’re shipping quickly across a wide range… testing 3.5… engaged with the U.S. government…”
Standard corporate deflection.
But let’s look past the words. One month is nothing in most industries. In AI? One month is an epoch. While Google polishes the chrome on Gemini 3.5 Pro, the competition just parked new cars on its lawn. Meta released a model that outruns Google’s current offerings. Meta beat them. On paper, anyway.
Two things are choking the release. First is the bloat. Google is huge. So many products integrate with Gemini that untangling the knots takes time. Startups don’t have this problem. They just launch. Google has committees.
Second? Fear.
Leaders at Google think the model isn’t competitive yet. That’s the stinging part.
Who is beating Google in the AI speed race?
The bar has moved since I/O.
Anthropic dropped Claude Mythos Preview. They called it advanced for cybersecurity, initially gatekeeping it for partners only. By June 9, they released a public-adjacent version called Fable 5.
Then came OpenAI. July 9 saw the announcement of GPT‑5.6 Sol. Coding capabilities? Sharp. Cybersecurity smarts? Also there.
And then… China enters the chat.
A lab called Moonshot released Kimi K3 this week. It’s open-source. Massive. 2.8 trillion parameters, for those keeping score. Testers say it rivals Fable 5 and GPT‑5.6 in capability.
But here is the kicker: it costs significantly less.
Google sits on a data mountain. They have direct lines to billions of Android phones. They should be crushing the curve. Instead, their rankings on the AI leaderboards are slipping. Without a frontier model to show the crowd, the applause fades.
What should users do while waiting for Gemini 3.5 Pro
If you are waiting for the next big thing from Google, patience is a virtue you didn’t choose. You’re just stuck with it.
Gemini 3.5 might launch soon. Pichai promised June. July came. Maybe August.
While Google plays it safe, refining code and negotiating frameworks, Anthropic and OpenAI are busy selling tickets to the next train. Kimi K3 is showing that performance doesn’t always have to come with a premium price tag or a Silicon Valley zip code.
Google has the infrastructure. It has the user base.
Does it have the speed?
Right now, the competitors say no. And until Gemini 3.5 Pro actually appears on the dashboard, nobody can prove them wrong.
Are you sticking with Flash or waiting for Pro?
Light Speed Update
Want more out-of-this-world tech news? Mashable’s weekly Light Speed newsletter covers the chaos. If you are 16+, sign up. Read the terms. They exist for a reason. 🚀
