Gemini 3.5 Flash – Built for Agents, Speed, and Scale

Trusted by users from 10,000+ companies

Gemini 3.5 Flash’s Core Capabilities

Gemini 3.5 Flash brings practical capabilities built for developers and enterprise teams deploying at scale.

Context Caching

Supported for long-context and repeated-query workflows, significantly reducing cost on sessions that reuse the same large inputs across multiple calls.

Thinking Mode

Built-in thinking support for complex reasoning tasks. Toggle on for hard problems, off for speed-sensitive, high-volume requests.

Google Search Grounding

Native integration with Google Search as a tool, enabling agents to retrieve real-time information within a single workflow.

Frequently Asked Questions

Learn more about Gooogle's newest, fastest, and most, cost-effective model through other people's questions..

Google Maps Grounding

Location-aware reasoning built in, enabling agents to incorporate geographic context directly into analysis and decision tasks.

Code Execution

Runs code natively within the model, supporting iterative development loops that generate, test, and refine within a single agentic pass.

Batch API

Full batch processing support for high-volume, asynchronous workloads. Process large queues of requests at lower cost without real-time latency requirements.

Flex and Priority Inference

Two inference tiers let teams balance speed and cost. Priority for time-sensitive requests, flex for throughput-optimised background tasks.

Life Sciences Precision

Box reported 96.4% greater accuracy on life sciences data extraction tasks versus the previous Flash generation, a meaningful improvement.

Android Studio Integration

Available natively inside Android Studio, giving mobile developers access to Gemini 3.5 Flash's coding and reasoning capabilities directly within their build environment.

Gemini 3.5 Flash Leads in Crucial Benchmarks

Gemini 3.5 Flash tops the benchmarks that reflect how agents, analysts, and developers actually work.

Best Multi-step Agent

Gemini 3.5 Flash scores 83.6% on MCP Atlas, the highest of every model in the comparison. It leads on the benchmark designed specifically to measure performance on real, multi-step workflows using tools.

Top Financial Reasoning Model

Gemini 3.5 Flash is ahead of every other model tested on Finance Agent v2. For financial analysis, structured reporting, and data-driven decision tasks, it consistently outperforms both faster and heavier alternatives.

Reasoning from Complex Charts

Gemini 3.5 Flash scores 84.2% on CharXiv, leading all models tested on synthesising information from complex visual data. For work that depends on reading charts, diagrams, and structured visuals accurately, it never misses.

Financial Reports Built from Raw Data

Financial services teams using Gemini 3.5 Flash report 46.7% greater accuracy in their reports than previous Flash models. It handles the full pipeline from reading source data to producing structured to decision-ready outputs.

Simultaneous Work Across Every Input Type

Gemini 3.5 Flash reasons across text, images, video, audio, and PDFs in a single session, topping multimodal understanding benchmarks. Analysts and researchers no longer need to route different inputs through different tools.

Complex Agentic Tasks, Run at Volume

Gemini 3.5 Flash is designed to run as a sub-agent inside larger workflows, handling iterative, multi-step tasks at Flash-level speed and cost. Armadin found it completed long-range, multi-turn tasks 42% more effectively.

Gemini 3.5 Flash – Frontier Intelligence at Flash Speed

Gemini 3.5 Flash’s Core Capabilities

Context Caching

Thinking Mode

Google Search Grounding

Frequently Asked Questions

What is Gemini 3.5 Flash?

How does Gemini 3.5 Flash differ from Gemini 3 Flash?

What is Gemini 3.5 Flash's context window?

What inputs does Gemini 3.5 Flash accept?

Does Gemini 3.5 Flash support thinking?

Where is Gemini 3.5 Flash available?

How does Gemini 3.5 Flash compare to Gemini 3.1 Pro?

Is Gemini 3.5 Flash suitable for enterprise use?

What is Gemini 3.5 Flash's knowledge cutoff?