Cortex — Veille numérique

Gemini

Gemini 3.1 Flash-Lite: Built for intelligence at scale

Introducing Gemini 3.1 Flash-Lite, our fastest and most cost-efficient Gemini 3 series model. Built for high-volume developer workloads at scale, 3.1 Flash-Lite delivers high quality for its price and model tier. Gemini 3.1 Flash-Lite can tackle tasks at scale, like high-volume translation and content moderation, where cost is a priority. And it can also handle more complex workloads where more in-depth reasoning is needed, like generating user interfaces and dashboards, creating simulations or

13 juil.

Nano Banana 2: Combining Pro capabilities with lightning-fast speed

Our latest image generation model offers advanced world knowledge, production-ready specs, subject consistency and more, all at Flash speed.

13 juil.

Gemini

A new era of intelligence with Gemini 3

Learn, plan, and build with Gemini 3. Available today, you can access Gemini 3 on Vertex AI and Gemini Enterprise.

13 juil.

A benchmark we trusted turned out to be broken

The field is converging on meta-benchmarking and exerting greater control of how we challenge and evaluate our agents. A growing body of work uses item response theory, the latent-ability model behind standardized testing, to treat difficulty as a measured quantity rather than a label: tinyBenchmarks and metabench show that a handful of informative items reproduce a model's full score, and PSN-IRT turns the same lens on benchmark quality itself. Others audit the ground truth directly: MMLU-Redux

13 juil.

When two maps disagree

Now the recursive turn: If difficulty is something we generate, then we need to evaluate the generator itself; we should not trust it blindly either. So we built the same ambiguity sweep two ways: steering terms from a pure-LLM guess, versus terms grounded in TF-IDF surprisal. The two disagreed violently. At high ambiguity, the LLM-built sweep scored the agent at F1 ≈ 0.34; the grounded sweep, ≈ 0.85. One of these maps is badly distorted. The grounded one, predictably, is the more robust: surpri

13 juil.

Gemini

Evaluate your evals

The versatility of large language models (LLMs) led to the creation of diverse benchmarks that thoroughly test a variety of language models' abilities. These benchmarks consist of tens of thousands of examples making evaluation of LLMs very expensive. In this paper, we investigate strategies to reduce the number of evaluations needed to assess the performance of an LLM on several key benchmarks. For example, we show that to accurately estimate the performance of an LLM on MMLU, a popular multiple-choice QA benchmark consisting of 14K examples, it is sufficient to evaluate this LLM on 100 curated examples. We release evaluation tools and tiny versions of popular benchmarks: Open LLM Leaderboard, MMLU, HELM, and AlpacaEval 2.0. Our empirical analysis demonstrates that these tools and tiny benchmarks are sufficient to reliably and efficiently reproduce the original evaluation results.

13 juil.

ADK and ComputeSDK built-in support

Build powerful multi-agent systems with Agent Development Kit (ADK)

13 juil.

Google

What’s new: Scaling massive data layers with C4N

Review network bandwidth limits and enable higher network bandwidths using gvNIC and per VM Tier_1 networking performance for Google Cloud.

13 juil.

Google

A deeper look at C4N shapes and specs

Thank you for your interest in participating in Google Cloud's preview program for C4N Local SSD VMs, which adds -lssd shapes to the C4N machines series. Preview capacity is available in us-central1, us-east1, us-west1 and europe-west2 regions and requires the customer's projects to be allowlisted. For more details on C4N machine series, see here. This preview period is for customers who are interested in testing performance and collaborating in our go-to-market campaigns. We are also looking for participation from our customers and partners to support our launch communications (e.g., providing a quote, permissions to use their logo etc.) - please reach out if interested. This form is to indicate that your customer is interested in participating in C4N Local SSD Preview. We will contact you directly once you submit a response with more details.

13 juil.

Gemini

Building on the infrastructure foundation powering Gemini

Enterprise customer experiences are entering a new era. Organizations are moving beyond traditional chatbots toward AI agents that can understand customer intent, reason across enterprise knowledge, and take action across business systems. As these experiences move into production, enterprises need more than powerful models. They need an AI platform that combines frontier research with enterprise security, governance, operational reliability, and the ability to scale globally. Today, Gemini Ente

13 juil.

Google

Moving beyond the exam

Learn and earn with Google Skills, a platform that provides free training and certifications for Google Cloud partners and beginners. Explore now.

13 juil.

Google

New ways to keep Google Cloud certifications current and boost your career

Google Cloud technology is ever-evolving. Refresh your knowledge and skills by renewing your cloud certifications. Renewing your certification maintains your certification status and certificate numbe

13 juil.

Intelligence Artificielle

The New Security Control Point: Governing AI Agents Inside the Execution Loop

Stop AI agents from going rogue with real-time runtime monitoring. Capsule Security protects AI agents across workflows with instant anomaly detection.

13 juil.

Cybersécurité

Application security and vulnerability management

Fix what matters. Once. aisy helps security teams turn vulnerability noise into threat models they can act on, so every finding makes the next one less likely.

13 juil.

Développement

Cloud, network, and infrastructure security

CloudFence uses network and identity logs to baseline behavior per workload, helping security teams quickly detect and respond to suspicious activities.

13 juil.

Endpoint security and data protection

Bold runs AI locally on every endpoint to stop data exfiltration, shadow AI risk, insider threats, and misdelivered emails — in real time, no cloud cost.

13 juil.

Cybersécurité

SOC automation and offensive security

COGNNA’s AI SOC platform with intelligent SOC Agents streamlines detection, response and compliance supporting SOC teams for faster, smarter security.

13 juil.

Security infrastructure and specialized services

From device management to threat protection, Netsec owns the full IT and cybersecurity lifecycle. Operated inside Slack or Teams.

13 juil.

Grand public numérique

Step 4: Publish your agent listing on Marketplace

Join the Google Cloud Partner Network: If you're new to offering your solutions on Marketplace, join the Google Cloud Partner Network. Review Agent-as-a-Service listing requirements. Verify that your organization meets the requirements to list your solutions on Marketplace. Marketplace Vendor Agreement: Review and accept the Marketplace Vendor Agreement (MVA). Nominate your agent for Google Cloud Marketplace by contacting your Google Cloud representative. All agents listed on Marketplace

13 juil.

Grand public numérique

Step 5: Managing transactions and registrations in Marketplace and the Gemini Enterprise App

Discover how Google Cloud's billing access control resource allows users to create & update access permissions for various roles in an organization.

13 juil.

Cortex — actualités tech et cybersécurité

lundi 13 juillet

Gemini 3.1 Flash-Lite: Built for intelligence at scale

Nano Banana 2: Combining Pro capabilities with lightning-fast speed

A new era of intelligence with Gemini 3

A benchmark we trusted turned out to be broken

When two maps disagree

Evaluate your evals

ADK and ComputeSDK built-in support

What’s new: Scaling massive data layers with C4N

A deeper look at C4N shapes and specs

Building on the infrastructure foundation powering Gemini

Moving beyond the exam

New ways to keep Google Cloud certifications current and boost your career

The New Security Control Point: Governing AI Agents Inside the Execution Loop

Application security and vulnerability management

Cloud, network, and infrastructure security

Endpoint security and data protection

SOC automation and offensive security

Security infrastructure and specialized services

Step 4: Publish your agent listing on Marketplace

Step 5: Managing transactions and registrations in Marketplace and the Gemini Enterprise App