Gemini 3.5: frontier intelligence with action - Google Blog
At Google I/O we released Gemini 3.5, our latest series of models combining frontier intelligence with action.
blog.googleHere’s the latest on Gemini 3.5 Flash and related 3.5 developments as of today.
Direct answer
- Gemini 3.5 Flash is the speed-focused variant in Google’s Gemini 3.5 lineup, introduced at Google I/O 2026. It’s designed to be fast, efficient, and capable in coding, agentic tasks, and multimodal workloads, and is set as the default model in the Gemini app and AI mode in Search for broad user access.[2][3][6]
Key details and what’s new
- Speed and latency: Gemini 3.5 Flash is promoted as significantly faster than earlier models, with claims of higher throughput (tokens per second) and lower latency, aiming to reduce the trade-off between reasoning power and responsiveness.[1][3][6]
- Use cases and features: It emphasizes improved coding capabilities, enhanced agent support, and efficient token usage. Google positions 3.5 Flash as the default consumer model in the Gemini app and in Search AI Mode, enabling everyday tasks with faster responses.[3][5][6]
- Availability and rollout: Gemini 3.5 Flash was rolled out to users via the Gemini app and Google AI services at launch, with broader availability planned through Google AI Studio APIs and developer channels. Google also announced a higher-end 3.5 Pro variant set to launch publicly in the following month(s) for more demanding workloads.[5][6][3]
- Benchmark and comparisons: Reports indicate 3.5 Flash outperforms Gemini 3.1 Pro on several benchmarks and approaches the capabilities of higher-end models while delivering faster output. Independent outlets highlight its strong performance in coding and multimodal tasks relative to prior generations.[6][1][2]
Context and related ecosystem
- Gemini Spark: Google described Gemini Spark as a personal AI agent framework powered by 3.5 Flash, rolling out to testers with broader beta plans for US users. This signals a push toward autonomous, action-oriented AI assistants within the Gemini ecosystem.[3]
- Market positioning: The 3.5 Flash line continues Google’s strategy of offering a fast, cost-effective “Flash” tier alongside more capable Pro and Ultra tiers, balancing performance with price for developers and end users.[5][6]
If you’d like, I can pull out a concise comparison table (3.5 Flash vs 3.5 Pro vs 3.1 Pro) and summarize typical use-case suitability (coding, agents, multimodal tasks) based on the latest benchmarks and official notes. I can also track any new rollout updates or availability in the UK/Europe if you’re planning to try it locally.