Gemini 3.5 Flash-Lite

Gemini 3.5 Flash-Lite shipped alongside Gemini 3.6 Flash on July 21, 2026 as the fastest model in the 3.5 series. Pricing is $0.30 per million input tokens and $2.50 per million output, and Artificial Analysis clocks it at 350 output tokens per second. Google positions it for high-throughput production traffic: agentic search, document processing, and any workload where latency and volume matter more than frontier reasoning. The gains over Gemini 3.1 Flash-Lite are large on Google's own numbers: Terminal-Bench 2.1 at 54 percent against 31, GDM-MRCR v2 long context at 72.2 percent against 60.1, and GDPval-AA v2 at 1140 against 642. More interesting for anyone still on an older Flash tier, Google reports it beating Gemini 3 Flash on SWE-Bench Pro (54.2 percent against 49.6) and OSWorld-Verified (74.0 against 65.1), which makes it a faster and cheaper option than the Flash model it sits under. Thinking levels are configurable, so the same model can run minimal-thinking high-volume batches or engage higher effort for multi-step subagent work, and computer use is a built-in tool. It is rolling out in Google Search as well as the API.