Google launches Gemini 3.8 Flash for coding and cybersecurity

Google launched two AI models on Wednesday: Gemini 3.8 Flash for software engineering and autonomous tasks, and Gemini 3.8 Flash Cyber for cybersecurity, marking its third Flash-generation release this quarter. Gemini 3.8 Flash is priced at $0.75 per million input tokens and $3.75 per million output tokens, matching its predecessor's introductory rate. It outperforms larger frontier models on the DeepSWE v1.1 benchmark while costing less, and scores 54.9% on HLE-Verified. Google attributes gains to architectural changes enabling deeper reasoning, though higher effort settings increase token consumption. The Cyber variant, available only through Google's new Fairwind Program for government and critical infrastructure partners, achieves a 47.2% pass@1 rate on CWE-Bench patching. Google reports internal testing showed it generated correct patches 2.6 times more often than larger commercial rivals and found a critical vulnerability in under two hours.