Urgent.News

What's breaking now, across thousands of outlets.

AI

Introducing Gemini 4 Neon

Google has unveiled Gemini 4 Neon, a cutting-edge AI model that packs frontier-level intelligence, capable of handling complex software engineering, knowledge work, and scientific workflows with remarkable speed. With a score of 61 on the Artificial Analysis Intelligence Index, Neon edges out comparable models and ranks #3 overall.

On the Vals Index, Neon claims the #2 spot and sits on the Pareto frontier. Neon outshines in steerability, ranking #2 in the Agent Arena Index, and excels in multi-step instruction following. It also boasts the lowest fabrication rate ever recorded in Google's internal Antigravity testing. Priced at $3 per million input tokens and $15 per million output tokens, Neon features a 1 million-token context window and is accessible in Google AI Studio, Gemini for business, Google Antigravity, and the Gemini app for Pro and Ultra subscribers.

Independent evaluations hail Neon as one of the top models, ranking #2 on the Vals Index, #1 on the Legal Research Bench and Tax Agent Bench, and #1 on Surge AI for finance, accounting, spreadsheet modeling, and complex constraint following. Neon also claims a 73.5% score on Chartography for reading dense professional charts, and matches leading models on Agents Last Exam v1.1 (75.8%).

Neon's performance in scientific and mathematical reasoning is top-tier, achieving the best in class on Terminal-Bench Science (70.5%), RiemannBench (84.0%), and HLE-Diamond (71.3%). Engineered for long-running agent workflows, Neon sustains performance across complex, tool-heavy tasks, whether running independently or as a high-judgment Orchestrator directing parallel Hydrogen subagents.

Its four levels allow teams to scale intelligence and adjust token spend per workload, delivering high performance at a lower cost per task compared to leading models on third-party benchmarks. Neon also excels in cybersecurity, generating more secure code than Carbon across all challenge tiers with half the cost. In production, Neon offers frontier-level performance across vulnerability finding, validating, and patching, all while being faster and more cost-efficient than large frontier models.

Written by urgent.news from Google DeepMind's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Every outlet that reported this, and when →

Read the original at deepmind.google →

More in AI

Managing AI Chat Context Without Sending the Entire Conversation

On a chat screen, you type a message and get a reply. Keep talking, and it feels like one continuous conversation. Behind that screen, your app sends your message to a backend.

  • Manage AI chat context by sending recent messages and summarizing older ones
  • Grow and flush approach combines sliding window with rolling summary
  • Keeps recent detail, compresses history, prevents indefinite growth

PlaceBench: we tested six AI assistants on 150 questions about real places

Ask an AI assistant "is anywhere near Slussen still serving hot food at 2 am on Monday?" and you get a confident answer. Is it right?

  • PlaceBench tested six AI assistants on 150 questions about real places in 121 cities.
  • Dots AI assistant scored highest at 92%, with no wrong answers.
  • Study found 12 closed places recommended, checked against public sources.

Chargebee Launches AI Agents That Help Audit Contracts and Collect Cash

Chargebee, a provider of monetization and billing infrastructure, has launched a new agentic finance platform, the company said in a Wednesday (Nov. 18) press release. At launch, the Arc platform’s artificial intelligence agents solve three jobs, including contract ingestion, receivables and revenue recognition, according to the release.

More from Wednesday 18 November →