Google Releases Gemini 3.8 Flash, Its Third Flash Model in Six Weeks
Google says the new model outperforms OpenAI and Anthropic rivals on complex engineering and agentic coding benchmarks while improving cyber defense tools.
- On Wednesday, Google launched Gemini Flash 3.8 and Flash Cyber models, expanding its current lineup after a three-week gap since the previous release.
- This marks Google's third Flash release in six weeks, with the new model often outperforming the three-week-old Gemini 3.7 Flash, particularly on agentic coding tasks.
- On Terminal-Bench 4.0, Flash 3.8 scores 19.1%, while Fable 5.1 hits 55.8%; Google describes the new release as its "workhorse model."
- Through the Fairwind Program, Google provides 650 partners access to autonomous tools; Four Flynn, VP for Security and Privacy, noted the edge comes from "shrinking the time between detecting a flaw and patching it."
- Google ships 3.8 Flash with safeguards per its Frontier Safety Framework; introductory pricing of $0.75/$3.75 per million tokens expires December 31, 2026, rising to $1.50/$7.50.
71 Articles
71 Articles
Google's Gemini 3.8 Flash Signals Relentless Pace in AI Race
Google just dropped another model. Three weeks after Gemini 3.7 Flash came out, the company unveiled Gemini 3.8 Flash on September 2. And a specialized version for cybersecurity. The message feels clear. Don’t count us out. The new offering arrives as questions swirl about Google’s flagship efforts. A full Gemini 3.5 Pro model has yet to appear despite earlier promises. Yet the Flash series keeps marching forward. Faster iterations. Focused gain…
(San Francisco = Yonhap News) Correspondent Kwon Young-jeon = Google has entered a race against time by releasing a new lightweight artificial intelligence (AI) model in just three weeks, but... unveiled on the same day...
Google releases Gemini 3.8 Flash and a cybersecurity variant limited to governments
Google has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, its third Flash release in six weeks, with the cyber model restricted to trusted testers and governments. Under the EU AI Act every general purpose model placed on the market carries documentation, copyright and training data obligations, and a systemic risk model must be […] This story continues at The Next Web
Coverage Details
Bias Distribution
- 46% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
