Skip to content
Google· Launch

Google releases three new Gemini models and begins training Gemini 4

Google DeepMind has unveiled three new Gemini models focused on high speed and lower token consumption, while the training of Gemini 4 has officially commenced.

By the Aheadline editorial team·13 okt. 2026·2 min read·Source: Entity-watch: Google DeepMindVerifierad signalAI-generated
Google releases three new Gemini models and begins training Gemini 4
Google releases three new Gemini models and begins training Gemini 4
Google releases three new Gemini models and begins training Gemini 4
By · Verktygs- & infrastrukturreporter
Last updated
Vad betyder det för mig?

What happened?

On 12 October 2026, Google DeepMind launched three new AI models in the Gemini family: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber. The flagship Gemini 3.6 Flash reduces output token usage by 17 percent compared to Gemini 3.5 Flash and achieves up to 65 percent token savings on the DeepSWE coding benchmark. Gemini 3.5 Flash Cyber is specifically optimised for cybersecurity applications, while Flash-Lite focuses on maximum speed and reduced costs.

Key facts

Lanseringsdatum12 oktober 2026
Tokenbesparing DeepSWE65 % minskning av utdatastokens
Generell tokenreduktion17 % färre tokens än 3.5 Flash
Nya modellerGemini 3.6 Flash, 3.5 Flash-Lite, 3.5 Cyber

Why it matters

The reduction in token counts and reasoning steps translates directly into lower operating costs and decreased latency for complex AI systems. By offering specialised variants for coding and cybersecurity, Google is addressing the demand for more cost-effective AI agents in production environments.

Who is affected?

The launch primarily concerns software developers, cybersecurity experts, and companies building large-scale AI agents. The lower token costs will benefit organisations running cost-sensitive applications in production.

Impact on the EU

The models are being released globally and are expected to be available to EU-based developers via Google AI Studio and Vertex AI. Existing privacy requirements under the EU GDPR and regulations within the EU AI Act apply to the deployment.

What else you should know

At the same time, Google DeepMind confirmed that it has begun large-scale pre-training of the next-generation flagship model, Gemini 4. The data regarding token savings is based on measurements from the evaluation platform Artificial Analysis Index.

Frequently asked questions

Quick answers about this story

Vad har hänt?
Google DeepMind lanserade tre nya AI-modeller: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite och Gemini 3.5 Flash Cyber, samt bekräftade förträningen av Gemini 4.
När hände det?
Lanseringen och tillkännagivandet gjordes den 12 oktober 2026.
Varför spelar det roll?
Modellerna erbjuder upp till 65 procents tokenbesparing vid kodning och lägre latens, vilket sänker driftkostnaderna för AI-agenter i produktion.
Vilka berörs av lanseringen?
Modellerna riktar sig primärt till mjukvaruutvecklare och företag som bygger automatiserade AI-agenter samt cybersäkerhetsverktyg.
Original source
Entity-watch: Google DeepMind·besthub.dev

The link opens in a new window and leads to the publisher's own site.

Verifierad signal

Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.

AI-verktyg i artikeln

Topics

#Google DeepMind#Gemini#Gemini 3.5 Flash#Agents#Models
[ STAY UP TO DATE ]

Get similar news straight to your inbox

No affiliate linksCancel anytimeGDPR-friendly
[ Frequency ]
[ What do you want to read about? ]

You'll receive updates on 2 topics.

The reader's room

Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.

Sign in to submit a comment or question.

Loading comments…
How this affects you

Read the article through your role

  • Decide whether this affects strategy over 6–12 months or is just noise.
  • Discuss with leadership: do we own the right question or does ownership need to move?
  • Ask: what risk are we taking by NOT acting on this this quarter?

Generated angle — not editorial analysis of "Google releases three new Gemini models and begins training "