Google launches Gemini 3.8 Flash and Gemini 3.8 Flash Cyber models
Google released Gemini 3.8 Flash for coding and agentic workflows, keeping the introductory price of $0.75 per million input tokens and $3.75 per million output tokens while improving from 65.3% to 73.7% on the DeepSWE benchmark. Google also launched a Gemini 3.8 Flash Cyber model for vulnerability discovery and patching, which scored 86.2% on the CyberGym vulnerability benchmark and produced 2.6x more correct patches than larger commercial models.
Detected & updated continuously · Source: Nebula
Story subjects
Sources
@markfenner
Another Google update?! Gemini 3.8 Flash is here. It keeps 3.7’s introductory price of $0.75 per million input tokens and $3.75 per million output tokens, while jumping from 65.3% to 73.7% on DeepSWE. That puts it within 0.3 points of Claude Opus 5 at a fraction of the listed https://t.co/0a37m4t29O
@techAU
A new day and new models. Google releases Gemini 3.8 Flash and Flash Cyber.
@StockSavvyShay
$GOOGL launched Gemini 3.8 Flash for coding and agentic workflows alongside a Cyber model built specifically for vulnerability discovery and patching. What stands out is the efficiency since cyber produced 2.6x more correct patches than larger commercial models. https://t.co/GnVU58huXG
@blckchaindaily
🚨 GOOGLE $GOOGL RELEASES GEMINI 3.8 FLASH CYBER, SCORES 86.2% ON CYBERGYM VULNERABILITY BENCHMARK
@CoinwatcherNews
🗞️ 🌐 #GOOGL lanzó Gemini 3.8 Flash para programación y flujos de trabajo agenticos junto con un modelo Cyber construido específicamente para el descubrimiento y parcheo de vulnerabilidades. Lo que destaca es la eficiencia, ya que cyber produjo 2.6x más parches correctos que https://t.co/Eg6p0R11Q0
@RoundtableSpace
Gemini 3.8 Flash drops today - codenamed Skimaki, on par with Opus 5 on coding according to WSJ. Google's engineers preferred it over Opus in internal tests. Fable 5.1 just dropped this morning. Google isn't waiting. https://t.co/qk54NTv0wY