Cover art for Google Unveils Gemini 3.8 Flash and Flash Cyber

Google Unveils Gemini 3.8 Flash and Flash Cyber

Google's newest AI models bring advanced agentic reasoning and automated cyber defense at low-cost Flash pricing.

8 cards · 1 min · tap to begin

From · · 1 min

Google Unveils Gemini 3.8 Flash and Flash Cyber

Google's newest AI models bring advanced agentic reasoning and automated cyber defense at low-cost Flash pricing.

In brief

Google's newest AI models bring advanced agentic reasoning and automated cyber defense at low-cost Flash pricing. Gemini 3.8 Flash brings frontier-level coding and cybersecurity defense capabilities to developer workflows without increasing token costs. Originally reported by Google.

Twin models for agents and defense

Google launched Gemini 3.8 Flash for general agentic workflows alongside Gemini 3.8 Flash Cyber, a specialized variant tuned for digital defense.

Our newest Gemini models deliver next-generation intelligence for agentic workflows and cybersecurity.

Frontier intelligence without added cost

The 3.8 Flash model maintains the low introductory price of 3.7 Flash at $0.75 per million input tokens and $3.75 per million output tokens despite substantial intelligence gains.

Gemini 3.8 introduces 2 variants... available at the same introductory price as 3.7 Flash.

Mastering autonomous software engineering

On long-horizon coding benchmarks like DeepSWE v1.1, 3.8 Flash outperforms larger frontier models in solving complex software tasks from start to finish.

3.8 Flash outperforms most larger frontier models in autonomously solving complex engineering problems end to end...

Adaptive reasoning that works harder

The model dynamically executes extra reasoning steps and calls tools iteratively on complex prompts. Developers can adjust effort levels to balance token usage with speed.

On complex tasks, it exhibits greater diligence — executing extra reasoning steps, and calling tools iteratively.

Targeted vulnerability discovery

Gemini 3.8 Flash Cyber achieves over a 70% success rate discovering vulnerabilities across codebases spanning 20 programming languages.

...the model showcases an impressive leap over our previous models and reaches a success rate exceeding 70%.

Automated patching for defenders

Google prioritized defense over offensive exploitation, delivering automated code patching performance on CWE-Bench that matches leading frontier models at far lower cost.

...we have invested in vulnerability fixing from the start, and prioritized it over offensive capabilities like exploitation.

Proven real-world security impact

Google's Cloud Vulnerability Research team used Flash Cyber to identify a critical security vulnerability in less than two hours, a process that traditionally takes months.

...found a critical foundational vulnerability in less than 2 hours, a vulnerability for which research and discovery usually takes months.

Controlled rollout via Fairwind Program

While standard 3.8 Flash is widely available to developers and consumers, the Cyber variant is restricted to trusted defenders and critical infrastructure operators.

...we’re providing trusted government authorities, as well as critical infrastructure operators and software maintainers with prioritized access...

The one thing

Gemini 3.8 Flash brings frontier-level coding and cybersecurity defense capabilities to developer workflows without increasing token costs.

Read the original on Google

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on ai and related topics.