Google Unveils Trio of New Models in the Gemini 3.5 Series, Alongside Exciting Updates on Gemini 3.5 Pro

Summary

• Google has unveiled three new models: Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber in CodeMender.
• The latest Gemini 3.6 Flash claims to lower output token usage by 17% compared to its predecessor.
• The company also stated that it would soon make Gemini 3.5 Pro widely accessible and has begun pre-training for Gemini 4.

In May, Google launched Gemini 3.5 Flash, touted as its most advanced speed model. Just two months later, the tech giant has introduced Gemini 3.6 Flash, alongside the 3.5 Flash-Lite and 3.5 Flash Cyber as part of its CodeMender initiative.

Gemini 3.6 Flash: Key Features

Gemini 3.6 Flash is presented as Google’s leading Flash model, enhancing coding capabilities, knowledge tasks, and overall performance. A significant highlight is its efficiency; data suggests that it utilizes 17% fewer output tokens compared to 3.5 Flash. The new model is also designed to require fewer reasoning steps and tool calls for complex workflows.

Moreover, this model is being offered at a lower cost than its predecessor. The pricing is set at $1.50 per million input tokens and $7.50 per million output tokens.

Performance improvements are notable; Google claims that 3.6 Flash exhibits greater accuracy, minimizes erroneous code edits, and decreases execution loops. The effectiveness of computer utilization has increased from 78.4% to 83%. Additionally, a new client-side tool has been integrated within the Gemini API and Gemini Enterprise, further enhancing user capabilities.

Gemini 3.5 Flash-Lite: The Fastest Model

Within the Gemini Flash series, 3.5 Flash-Lite is recognized for its speed and efficiency in low-latency scenarios. It operates at a rapid rate of 350 output tokens per second, with a cost of $0.30 per million input tokens and $2.50 per million output tokens. Apart from speed, it also boasts enhanced quality compared to 3.1 Flash-Lite, and it is reported to outperform even the previous 3 Flash model.

Gemini 3.5 Flash Cyber: Focusing on Cybersecurity

In contrast to the other models, Gemini 3.5 Flash Cyber is specifically designed for cybersecurity purposes. Built upon the foundation of 3.5 Flash, this model is optimized to identify and rectify vulnerabilities at a lower cost. Google mentions that when integrated into CodeMender, multiple 3.5 Flash Cyber agents will collaborate to generate a unified report. This model is slated for future limited-access trials specifically with governments and trusted partners.

Availability

Both the Gemini 3.6 Flash and 3.5 Flash-Lite are being rolled out now. Users can access them through the Gemini app, with 3.5 Flash-Lite also appearing in Search. Developers will find these models available via the Gemini API in Google AI Studio and Android Studio. For enterprise clients, access will be through the Gemini Enterprise Agent Platform.

In a recent update, Google disclosed that Gemini 3.5 Pro is making strides in testing with various partners, with plans for broad rollout on the horizon. Furthermore, the pre-training phase for the next significant model, Gemini 4, has commenced.

We appreciate your engagement with our content. Please review our Comment Policy before contributing.