Google Launches Gemini 3.8 Flash AI Model with Enhanced Performance but Potential Higher Costs
Google has launched its latest artificial intelligence model, Gemini 3.8 Flash, which it claims performs more reasoning steps and works harder than the previous version, Gemini 3.7 Flash. The new model maintains the same introductory pricing of $0.75 per million input tokens and $3.75 per million output tokens but may cost users more due to increased token usage for improved performance. Gemini 3.8 Flash outperforms its predecessor on software engineering benchmarks like DeepSWE v1.1, as well as in finance and legal agent evaluations such as Vals Finance Agent V2 and Harvey’s Legal Agent benchmark. The model also includes safeguards against misuse in sensitive areas including chemical, biological, radiological, nuclear (CBRN), and cyber offense domains. Alongside the new AI model, Google introduced Gemini 3.8 Flash Cyber under its Fairwind Program for governments and trusted partners like CrowdStrike and Center for Internet Security. The program also includes CodeMender, an agent designed to autonomously identify and fix vulnerabilities in critical infrastructure.
