Google launches three new Gemini models, skips flagship 3.5 Pro
Google's new Gemini lineup includes a dedicated cybersecurity model, but no Pro update
On Tuesday, Google DeepMind announced three new Gemini models: 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. Gemini 3.6 Flash is positioned as a workhorse model that delivers improved capabilities in coding, knowledge work, and multimodal performance while reducing token usage by up to 17%, making it cheaper than its predecessor. 3.5 Flash-Lite is billed as the most cost-effective model in its class, and 3.5 Flash Cyber is a specialized fine-tune for finding and fixing cybersecurity vulnerabilities, available exclusively to governments and trusted partners under a limited access pilot program. Google emphasized that these releases are about delivering efficiency, latency, and reliability to customers building AI agents at scale.
The launch notably lacks the expected Gemini 3.5 Pro, which was teased as a next-month release in May. Bloomberg recently reported internal delays as Google struggled to meet performance goals for the high-end model. In the same period, OpenAI launched GPT-5.5 and began rolling out GPT-5.6, while Anthropic released Claude Opus 4.8, Sonnet 5, and expanded access to Fable 5, highlighting the intense release pace from rival labs. Google DeepMind product lead Logan Kilpatrick said the company is testing Gemini 3.5 Pro with partners and hopes to ‘land soon,’ and that the team has started its most ambitious pre-training run yet for Gemini 4.
- Gemini 3.6 Flash reduces token usage by up to 17% while improving coding, knowledge, and multimodal capabilities.
- The new 3.5 Flash Cyber model is fine-tuned for cybersecurity vulnerabilities, available only to governments via a limited pilot program.
- Google's flagship Gemini 3.5 Pro remains delayed amid internal performance struggles, while OpenAI and Anthropic release competing high-end models.
Why It Matters
Google's focus on cheaper, specialized models signals a shift toward efficiency, but the missing Pro may delay high-end AI capabilities for enterprises.