Google DeepMind has postponed the launch of its flagship Gemini 3.5 Pro model to July 17, extending development to further improve performance before its public release. The delay comes as competition among leading AI companies intensifies, with Google refining its next-generation model to better compete with recent offerings from OpenAI and Anthropic.

According to reports, DeepMind decided to continue pre-training and optimization rather than releasing the model on its original timeline. The additional development period is expected to strengthen Gemini 3.5 Pro's capabilities in areas such as mathematical reasoning, SVG scene generation and image creation, addressing feedback received during internal testing and early evaluations.

The move follows a broader trend among AI developers prioritizing model quality over rapid deployment. As frontier AI systems become increasingly sophisticated, companies are spending more time refining reasoning accuracy, multimodal performance and reliability before expanding public access.

Google is also broadening its AI portfolio beyond Gemini 3.5 Pro. Reports indicate the company is preparing additional models, including Gemini 4 Flash and Nano Banana Pro, to serve different user groups ranging from developers and enterprises to everyday consumers. The strategy reflects Google's effort to provide specialized AI products for varying performance and pricing requirements.

Pricing is also expected to play an important role in Google's AI roadmap. Industry reports suggest DeepMind is focusing on delivering competitive performance while maintaining affordable usage costs, an approach aimed at attracting businesses and developers seeking alternatives to premium AI services.

The delay comes as OpenAI has begun rolling out its GPT-5.6 family, including Sol, Terra and Luna, while Anthropic continues expanding its Claude lineup. The increasingly crowded AI market has intensified the race to deliver faster, more capable and cost-effective models across coding, reasoning and multimodal tasks.

Despite the revised timeline, Google has not indicated that the delay reflects technical problems. Instead, the company appears to be using the additional development period to fine-tune the model before a broader rollout. Previous company statements have emphasized that Gemini 3.5 Pro remains one of DeepMind's highest-priority AI projects and will be released once it meets the company's performance and quality expectations.

With the July 17 launch window now in focus, Gemini 3.5 Pro is expected to become Google's flagship reasoning model, strengthening its position in an increasingly competitive artificial intelligence landscape.