AI News Summary — Week 11.2, 2026

This week's AI landscape brings major releases from OpenAI, Google, and Anthropic. Here's what's happening.

OpenAI Releases GPT-5.4 — Its "Most Capable" Model Yet

OpenAI has unveiled GPT-5.4, which the company describes as its most capable and efficient frontier model for professional work OpenAI. The release includes both GPT-5.4 Thinking (in ChatGPT) and GPT-5.4 Pro for maximum performance.

Key features:

• Native computer use: GPT-5.4 is OpenAI's first general-purpose model with state-of-the-art computer-use capabilities, enabling agents to operate computers and complete complex workflows across applications.
• 1 million token context: The model supports up to 1M tokens of context, allowing agents to plan, execute, and verify tasks across long horizons.
• Professional work improvements: On GDPval (testing agents' abilities across 44 occupations), GPT-5.4 matches or exceeds industry professionals in 83.0% of comparisons, up from 70.9% for GPT-5.2.
• Better spreadsheet and presentation work: On spreadsheet modeling tasks, GPT-5.4 achieves 87.3% mean score vs 68.4% for GPT-5.2. Human raters preferred GPT-5.4 presentations 68% of the time.
• Reduced hallucinations: GPT-5.4 is 33% less likely to produce false claims and 18% less likely to contain any errors compared to GPT-5.2.

The model also excels at computer use benchmarks, achieving 75% success on OSWorld-Verified (desktop navigation) — surpassing human performance at 72.4%.

Google Maps Gets "Ask Maps" — A Gemini-Powered Conversation AI

Google has announced a significant AI upgrade to Google Maps, introducing a Gemini-powered conversational feature called "Ask Maps" TechCrunch.

The feature lets users ask complex, real-world questions using natural language. Examples include:

• "My phone is dying, where can I charge it without having to wait in a long line for coffee?"
• "Is there a public tennis court with lights on that I can play at tonight?"
• "I'm headed to the Grand Canyon, Horseshoe Bend, and Coral Dunes, any recommended stops along the way?"

Ask Maps personalizes answers using signals including places users have searched for or saved. The feature is rolling out now in the U.S. and India on Android and iOS, with desktop coming soon.

Immersive Navigation update: Maps is also getting a complete redesign with 3D views showing buildings, overpasses, and terrain, plus enhanced road details like lanes, crosswalks, and traffic lights. Voice guidance has been updated to sound more natural, explaining trade-offs between routes (longer but less traffic vs faster with tolls).

Anthropic Launches Code Review Tool for AI-Generated Code

Anthropic has launched Code Review, a new product designed to check the flood of AI-generated code being produced by tools like Claude Code TechCrunch.

The tool addresses a growing problem: as "vibe coding" (using AI to generate code from plain language instructions) has accelerated development, it has also introduced new bugs, security risks, and poorly understood code.

How it works:

• Integrates with GitHub and automatically analyzes pull requests
• Leaves comments directly on code explaining potential issues and suggested fixes
• Focuses on logical errors (not style) to provide actionable feedback
• Uses color-coded severity labels: red (highest), yellow (potential issues), purple (historical bugs)
• Relies on multiple agents working in parallel, each examining code from different perspectives

The product is launching in research preview for Claude for Teams and Claude for Enterprise customers. Enterprise customers like Uber, Salesforce, and Accenture are already using Claude Code, which has seen subscriptions quadruple since the start of the year. Claude Code's run-rate revenue has surpassed $2.5 billion.

This roundup is brought to you by Conclio. Stay tuned for more AI news next week.