Gemini 3.5 Pro Reportedly Shelved as Google Shifts Focus to Gemini 4

Google’s long-awaited Gemini 3.5 Pro model is reportedly facing a major change in its development roadmap. A new institutional note from AI and semiconductor research firm Semi Analysis claims that Google has effectively shelved the model and is instead putting greater emphasis on Gemini 4.

However, this is not an official cancellation announcement from Google. Google has continued to indicate that Gemini 3.5 Pro is being tested, making the shelving claim a reported development rather than a confirmed company decision.

The uncertainty follows months of delays. Google had previously indicated that Gemini 3.5 Pro would arrive after the initial Gemini 3.5 releases, but the expected launch window passed without a public release.

As of early August, the model had no publicly confirmed release date or pricing information.

Months of Delays Raise Questions

The delay has become increasingly significant because Gemini 3.5 Pro was expected to serve as Google’s higher-end model in the Gemini 3.5 generation.

Instead, Google has continued expanding its Flash lineup while the Pro model remains unavailable.

According to the supplied report, the company launched Gemini 3.6 Flash, along with Gemini 3.5 Flash-Lite and the cybersecurity-focused Flash Cyber model.

At the same time, Google has confirmed that it has begun what it describes as its most ambitious pre-training run yet for Gemini 4.

That development has led Semi Analysis to interpret Google’s strategy as a shift away from Gemini 3.5 Pro and toward the next generation.

Gemini 4 Becomes Google’s Next Major Focus

Google’s work on Gemini 4 is now an important part of the company’s AI roadmap.

Starting a major pre-training run does not itself prove that Gemini 3.5 Pro has been cancelled. AI companies frequently develop multiple generations of models simultaneously.

Nevertheless, the combination of Gemini 3.5 Pro’s delays and Google’s emphasis on Gemini 4 has intensified speculation about the future of the unreleased model.

The latest reporting suggests that Gemini 4 could still be some time away, potentially leaving Google dependent on its existing Gemini models while competitors continue developing their own frontier systems.

Performance Concerns Add to the Pressure

The delay has also raised questions about whether Gemini 3.5 Pro would provide a sufficiently large improvement over Google’s existing models.

The supplied report says the model has appeared briefly on the LMSys Arena testing platform, where testers reportedly viewed its performance as incremental rather than a major generational advance.

Semi Analysis has also placed Google behind several competing AI labs on its Artificial Analysis Intelligence Index.

These rankings can change as new models are released and benchmarks are updated, so they should not be treated as definitive measures of overall AI capability.

Leadership Changes Add to the Uncertainty

The Gemini roadmap is also being reshaped during a major Google DeepMind leadership transition.

Demis Hassabis has stepped away from the day-to-day CEO role at Google DeepMind to become chairman of Google DeepMind and chief scientist for Alphabet. Koray Kavukcuoglu, previously Google DeepMind’s CTO, has taken on greater operational responsibility for the organisation and Gemini development.

Google chief scientist Jeff Dean has also left after 27 years to co-found Discovery Loop with other senior researchers.

The changes add another layer of uncertainty as Google attempts to accelerate its AI development and compete with companies such as OpenAI and Anthropic.

What the Gemini 3.5 Pro Delay Means

For developers, the biggest issue is uncertainty.

Teams waiting for Gemini 3.5 Pro cannot yet plan around a confirmed launch date, pricing structure or final capabilities. As a result, developers may need to continue using Google’s currently available Gemini models or consider alternatives while waiting for greater clarity.

If Google eventually moves directly toward Gemini 4, the company will need to explain how the new generation addresses the capabilities that developers expected from Gemini 3.5 Pro.

Frequently Asked Questions

Has Google cancelled Gemini 3.5 Pro?

Not officially. Semi Analysis has reportedly claimed that Google has shelved the model, but Google has not publicly confirmed its cancellation.

Why has Gemini 3.5 Pro been delayed?

The supplied report points to repeated delays and concerns around its performance, particularly in coding and agentic tasks. Google has not publicly confirmed that these issues are the reason for the delay.

Is Google already working on Gemini 4?

Yes. Google has confirmed that it has begun a major pre-training run for Gemini 4.

When will Gemini 4 launch?

No official general-availability date has been announced.

Conclusion

The future of Gemini 3.5 Pro remains uncertain after months of delays and a report from Semi Analysis claiming that Google has shifted its focus toward Gemini 4.

However, it would be misleading to describe the model as officially cancelled because Google has not confirmed that decision.

With Gemini 4 already entering a major pre-training phase and Google undergoing significant AI leadership changes, the company appears to be reorganising its strategy around its next generation of models. For developers, the immediate priority is to watch for an official Google announcement that clarifies whether Gemini 3.5 Pro will eventually launch or whether its capabilities will instead be incorporated into the company’s future Gemini roadmap.