"Performance is stunning," Google’s large model unusually builds hype before release—Gemini 3.0 to debut this week?

```

Google's upcoming Gemini 3.0 artificial intelligence model is attracting widespread attention across the industry.

Prediction markets show that the model will be unveiled next week, and CEO Sundar Pichai responded to related speculation on social media with a “thinking emoji,” almost confirming this timeline. This is Google's first time launching such a large-scale internal and external promotional campaign prior to releasing a large model.

Moreover, people who have interacted with the model rate its capabilities very highly. According to a Business Insider report on Monday, insiders described the new model as “extremely impressive,” and it is expected to make significant improvements in coding and multimedia content generation. Google employees have started expressing excitement about the launch on social media—something rarely seen before previous Google model releases.

Testing results in professional fields indicate breakthrough progress. Mark Humphries, a history professor at Wilfrid Laurier University in Canada, tested a suspected unreleased Gemini 3.0 model via Google AI Studio and found that its recognition of 18th-century handwritten manuscripts was nearly perfect, with a character error rate of just 0.56% and a word error rate of 1.22%, an improvement of 50%-70% over the previous Gemini 2.5 Pro—reaching expert human performance.

For Google, which fell into a “red alert” phase after ChatGPT’s release at the end of 2022, Gemini 3.0 is seen as a key step in reshaping its market position, especially as OpenAI’s much-anticipated ChatGPT-5 did not immediately have a major impact.

Rare Pre-Launch Hype

The atmosphere before Google's large model debut is markedly different this time. Prediction markets are betting that Gemini 3.0 will launch next week, and Pichai’s “thinking emoji” in response to discussions on social platform X has been widely interpreted as confirmation of the release time.

Google employees are unusually active on social media. Many have openly expressed excitement for the new model's launch, and this collective pre-release enthusiasm is rare for Google. Not only insiders, but many external individuals who have previewed the model's abilities have also posted passionate reviews online.

According to Business Insider, insiders have called the model “extremely impressive,” and it is expected to bring significant improvements to coding and multimedia content generation—possibly including major upgrades to Google's popular image tool NanoBanana.

Professional Testing Shows Breakthrough Abilities

Mark Humphries’ tests provide concrete cases for understanding the new model’s abilities. He used his professional work—analyzing 18th-century handwritten accounting ledgers—as benchmark tests. This task is extremely difficult, requiring identification of messy handwriting and integration of historical context, linguistic subtleties, and logical reasoning.

Humphries noted that deciphering historical handwritten texts requires abilities beyond visual recognition. “When you go back in time, you enter a different country. People speak differently, use unfamiliar words or familiar words in unfamiliar ways. Past people used different measurement and accounting systems, different wording, punctuation, capitalization, and spelling.”

Test results show that the previous Gemini 2.5 Pro had a character error rate of about 4% on these complex documents, roughly equivalent to a professional human transcriber. The new model reduced character error rate to 0.56% and word error rate to 1.22%, reaching expert human performance standards.

Even more noteworthy is the model’s reasoning ability. Humphries found the model could spontaneously engage in step-by-step symbolic reasoning, for example, deducing “145” in an 18th-century merchant ledger as “14 pounds 5 ounces”—not just text recognition, but understanding the economic and cultural systems behind these records.

A Turning Point in Google’s AI Strategy

For Google, the release of Gemini 3.0 is of strategic significance. Since the end of 2022 and the release of ChatGPT, Google has been seen as playing catch-up in the AI race and even issued an internal “red alert.” Business Insider quoted insiders as saying, the new model may give Google a chance to reclaim the lead, especially after OpenAI's ChatGPT-5 did not immediately generate a major impact.

The model is expected to deliver significant improvements in coding and multimedia content generation. Google's image generation model NanoBanana has recently received positive user feedback. The name comes from a placeholder created by an employee named Nina. According to Google Gemini App Product Manager David Sharon on the Made By Google podcast, Google used this name when anonymously submitting the model to the open AI evaluation platform LM Arena for fair testing—and it unexpectedly became popular in online communities, prompting Google to officially adopt the name.

The most profound meaning is that if the new model’s abilities are systematically validated, AI may be transitioning from the complex “stochastic parrot” to a system with true understanding. Humphries pointed out: “If such behavior proves reliable and reproducible, it points to something profound: true reasoning may not require explicit rules or a symbolic framework, but can emerge from scale, multimodality, and exposure to sufficient structured complexity.”

For historians, near-perfect handwritten text recognition combined with contextual understanding will allow for rapid digitization and analysis of centuries-trapped knowledge, possibly rewriting our understanding of the past. For broader applications, a reasoning-capable AI could begin to automate complex cognitive tasks that were previously considered the domain of human experts.

Risk Warning and DisclaimerMarkets entail risks, and investment requires caution. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article are applicable to their specific situation. Investing accordingly is at your own risk. ```