La stratégie d'optimisation des coûts d'inférence répond directement à la demande des entreprises qui déploient des agents à grande échelle. Le retard répété sur le modèle Pro indique toutefois des difficultés techniques à franchir un palier de raisonnement sans exploser les coûts de calcul.
Ce choix privilégie l'adoption industrielle immédiate en abaissant la facture de calcul pour les agents d'intelligence artificielle en production. Google concède du terrain sur le créneau des modèles généralistes à forte capacité, où la concurrence accentue la pression commerciale.
Google oriente ses sorties vers l'efficience économique et la vitesse avec trois déclinaisons légères. Gemini 3.6 Flash diminue la consommation de tokens de 17 % par rapport à la version 3.5 Flash tout en renforçant ses capacités en programmation. La variante Flash Cyber cible la détection de vulnérabilités informatiques et reste réservée aux gouvernements ainsi qu'aux partenaires agréés. L'absence de Gemini 3.5 Pro confirme des retards internes face aux objectifs de performance fixés par la filiale de Google.
Impact concret
Les équipes techniques bénéficient d'une réduction directe de 17 % sur les coûts de traitement des requêtes textuelles et visuelles. Les organisations publiques et de défense disposent d'un outil dédié à l'audit automatisé du code source contre les failles de sécurité.
Applications concrètes
Déploiement d'agents de code automatisés à bas coût et audit ciblé de failles logicielles pour les infrastructures critiques.
On Tuesday, Google DeepMind released Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. Gemini 3.6 Flash is Google’s “workhorse model” that promises improved capabilities in coding, knowledge work, and multimodal performance while reducing token usage by up to 17%, making it cheaper than its predecessor 3.5 Flash. Gemini 3.5 Flash-Lite is the most cost-effective model in the class, and 3.5 Flash Cyber is a specialized model that was fine-tuned for finding and fixing cybersecurity vulnerabilities at a decent price point. This model will be exclusively available to governments and trusted partners as part of a limited access pilot program, according to Google. Google says the focus on these releases is to deliver efficiency, latency, and reliability to customers that are building AI agents at scale. The launch is notable not just for what Google shipped — cheaper, faster models optimized for coding, efficiency, and cybersecurity — but also for what it didn’t. The update doesn’t include the long-anticipated update to Google’s flagship model, Gemini Pro, which was last updated in February. In the time since that launch, OpenAI has released GPT-5.5 and begun rolling out GPT-5.6 , while Anthropic has launched Claude Opus 4.8 and Claude Sonnet 5 and has expanded access to its frontier Fable 5 model , highlighting the intense release pace of the rival labs. Google teased the release of Pro as part of the 3.5 Flash release in May, saying the Pro version was “already being used internally, and we look forward to rolling it out next month.” Last week, Bloomberg reported that Google was facing internal delays in launching the 3.5 Pro as it struggled to meet internal performance goals. Gemini Pro models are generally Google’s highest-capability offerings for complex reasoning and coding tasks, while Flash models prioritize lower cost and faster response times for production applications. Google DeepMind product lead Logan Kilpatrick said Tuesday that the company is currently testing Gemini 3.5 Pro with partners and hopes to “land soon.” He also noted that the team has started its most ambitious pre-training run yet for Gemini 4. Topics AI , Enterprise , Google , google gemini When you purchase through links in our articles, we may earn a small commission . This doesn’t affect our editorial independence. Rebecca Bellan Senior Reporter Rebecca Bellan is a senior reporter at TechCrunch where she covers the business, policy, and emerging trends shaping artificial intelligence. Her work has also appeared in Forbes, Bloomberg, The Atlantic, The Daily Beast, and other publications. You can contact or verify outreach from Rebecca by emailing rebecca.bellan@techcrunch.com or via encrypted message at rebeccabellan.491 on Signal. View Bio October 13 – 15 San Francisco Scale faster. Grow your portfolio. Gain practical expertise. No matter your goal, Disrupt can empower you. Save up to $330 toda y! REGISTER NOW Most Popular OpenAI says Hugging Face was breached by its pre-release models Russell Brandom Jack Dorsey is taking on Slack with Buzz, a group chat platform for teams and their AI agents Amanda Silberling Light made a flip phone — it’s colorful and it’s cheap Amanda Silberling AI music generator Suno breach affects 55M users, per Have I Been Pwned Zack Whittaker Anthropic’s landmark $1.5B copyright settlement is approved Kirsten Korosec Google is working on a new AI chip designed to make Gemini more efficient Lucas Ropek Judge pauses $110B Paramount-Warner Bros. merger Aisha Malik