Google has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, marking its third Flash launch in six weeks. The company calls 3.8 its best reasoning and coding model yet, matching the speed and low cost of 3.7 Flash while approaching the performance of larger frontier models.
Gemini 3.8 Flash targets long-horizon software engineering and autonomous agents. On the DeepSWE v1.1 benchmark it solves complex engineering problems end to end, beating most larger frontier models at a fraction of the cost. Google says the gains come from the model working harder, running extra reasoning steps and iterating on tool calls, with adjustable effort levels for token-sensitive workloads.
Gemini 3.8 Flash Cyber, available to trusted defenders through the Fairwind Program, focuses on autonomous vulnerability discovery and automated patching. It scores a 47.2% pass@1 on CWE-Bench versus 47.8% for a leading frontier model at far lower cost, and Google already uses it to secure its own code. Introductory pricing runs until the end of 2026.