2026-W40 · 31 stories

The week in AI: September 28, 2026 – October 4, 2026

The week's main developments, grouped by original publication date. Every item links to our full story and the original source.

Issue date: ·Week: September 28, 2026 – October 4, 2026

What changed this week

During the period from September 28 to October 4, 2026, significant updates in AI development and application were introduced by GitHub Copilot, Salesforce, and Microsoft Foundry.

Three key changes

  1. GitHub Copilot: Deprecation of Four Models and Alternatives

    What changed
    As of October 2, 2026, four models (Gemini 3.5 Flash, Gemini 3.8 Flash, Gemini 3.6 Flash, Kimi K2.7 Code, Claude Opus 4.7) were deprecated across all Copilot experiences.
    Practical impact
    Users must review their workflows and integrations, while Copilot Enterprise administrators need to enable access to alternative models via model policies.
    What to consider
    Enterprise users require manual configuration to access alternatives through Copilot settings.
    Full story ↗
  2. Salesforce: AI Agents in Customer Service Examples

    What changed
    Companies utilize AI agents to autonomously resolve 89% of routine messaging inquiries and 15% of calls in customer service.
    Practical impact
    Examples illustrate how automation can integrate with staff preparation and model behavior control.
    What to consider
    Results are based on company self-reports and lack independent verification.
    Full story ↗
  3. Salesforce: New CRM Model Koa and Enterprise AI Harness

    What changed
    Salesforce introduced the CRM model Koa, built on NVIDIA Nemotron 3 Super architecture, improving context retention and response relevance in complex conversations.
    Practical impact
    Enterprise AI Harness provides six capabilities for AI governance: trusted models, context, agent functionality, actions, governance, and security.
    What to consider
    Access is limited to Salesforce platform users and developers.
    Full story ↗

Models

6

OpenAI Launches GPT-6.1 Sol Model With Reduced API Costs

OpenAI has introduced GPT-6.1 Sol, a cost-efficient model providing performance comparable to GPT-6 Astra for coding and professional tasks.

The release provides a more economically viable option for developers building complex AI agents and software engineering tools that require high reasoning capabilities.

Full story ↗Original source ↗

Anthropic: GLM-5.3 built exploits and demonstrated bypassable safeguards in simulations

Anthropic published an analysis of Zhipu AI’s GLM-5.3, showing its ability to develop exploits and vulnerabilities in its safeguards that can be bypassed in simulations without code execution.

Anthropic’s findings indicate that GLM-5.3, an open-weight model, can autonomously build exploits and has safeguards that are easily bypassed in simulations, posing significant cybersecurity risks for open AI models.

Full story ↗Original source ↗

Products

3

NVIDIA Releases Metropolis Blueprint VSS 3.3 Update

NVIDIA has released Metropolis Blueprint VSS 3.3, featuring a new vss-build-vision-ai skill for creating video analytics agents from text prompts.

This update streamlines the development of vision-based AI workflows by allowing engineers to generate functional agents through simple prompting rather than manual configuration.

Full story ↗Original source ↗

OpenAI introduces Dots, autonomous AI agents powered by GPT-6 Astra

OpenAI is rolling out Dots, persistent AI agents that manage third-party applications and perform proactive research for Pro and Enterprise users.

The launch of Dots shifts AI utility from interactive chat to persistent automation. By offloading multi-step tasks to agents that operate 24/7, users can minimize manual coordination between different software platforms.

Full story ↗Original source ↗

Applications

7

Salesforce shares examples of Agentforce in customer service

Canada Goose's AI agents autonomously resolved 89% of messaging inquiries and 15% of calls, improving service efficiency.

The examples show how service automation can involve staff preparation and controls on model behavior. The results are company-reported; the material provides no methodology or baselines and does not independently verify them.

Full story ↗Original source ↗

Barclays expands Claude use and expects Claude Code adoption to grow by 2027

Barclays is expanding Claude use across the bank. Anthropic describes an employee assistant and email processing; Barclays expects most of its software engineers to use Claude Code in 2027.

The announcement describes Claude use in software development and internal banking operations. Adoption and email-processing figures are reported by Anthropic; the Claude Code figures are Barclays’ expectations, not achieved results.

Full story ↗Original source ↗

Microsoft Research Develops AI for Forecasting Space-Weather Risks to Power Grids

A machine-learning system estimates space-weather risks at 66,935 substations in the continental US, producing location-specific forecasts 30–60 minutes before potential impacts.

After further validation with utilities and operational data, the estimates could help operators prioritize engineering reviews and consider targeted protective actions.

Full story ↗Original source ↗

Salesforce Launches Koa CRM Reasoning Model

Salesforce introduced the Koa CRM model built on NVIDIA Nemotron 3 architecture and a new Enterprise AI Harness for AI governance and trusted operations.

The platform enables organizations to select and govern specific AI models for business tasks, ensuring that AI agents remain connected to enterprise data and security policies.

Full story ↗Original source ↗

Developers

15

GitHub opens Copilot code review API requests to five plans; Balanced is the default

Users on Copilot Pro, Pro+, Max, Business, and Enterprise can request code reviews through supported REST and GraphQL APIs and optionally set the effort level for each request.

Teams on the listed plans can request reviews through the supported APIs from their own scripts, workflows, and internal tools. They can set the effort level per request, while the Balanced default leaves an explicitly selected Lite setting intact.

Full story ↗Original source ↗

GitHub has deprecated four models across Copilot experiences on October 2, 2026

The change covers every Copilot experience. GitHub suggests three alternatives; Copilot Enterprise administrators may need to enable access through model policies.

The change affects model choices across Copilot experiences. Users should review workflows and integrations, while Enterprise administrators may need to check access to alternatives in their settings.

Full story ↗Original source ↗

Microsoft introduces three MAI speech models in Microsoft Foundry

Microsoft Foundry offers streaming transcription across 60 languages and two speech-generation models. Microsoft has published prices and named platforms it plans to add.

Developers can pair streaming transcription, with first partials arriving within the low hundreds of milliseconds, with either of two text-to-speech models. Published per-hour and per-1M-character prices support cost planning. Microsoft’s performance claims have not been independently verified here.

Full story ↗Original source ↗

Google Cloud Data Agent Kit Now Generally Available

Google Cloud has released its Data Agent Kit, providing Model Context Protocol tools for integrating AI coding agents with cloud data services.

This toolkit enables developers to connect their existing coding agents directly to analytics and operational databases for more efficient data workflows.

Full story ↗Original source ↗

CoreWeave Adds NVIDIA Vera Rubin NVL72 Systems to Cloud

CoreWeave provides early-access to NVIDIA Vera Rubin NVL72 systems integrated with Spectrum-X networking in its cloud, alongside new tools for agentic AI workflows.

These updates are significant for AI engineering teams, as demonstrated by Cognition's performance gains in real-time code generation and responsive multistep reasoning.

Full story ↗Original source ↗

Hugging Face introduces Open TTS Leaderboard for objective model evaluation

Hugging Face’s Open TTS Leaderboard compares open TTS models using WER/CER, RTFx, TTFA and speaker similarity. Its scores complement rather than replace human preference rankings.

For developers, the leaderboard offers a faster and more reproducible way to compare models than collecting votes for weeks. Its metrics are still proxies: WER and speaker similarity do not measure naturalness, expressiveness or listener preference, and the leaderboard does not replace human rankings.

Full story ↗Original source ↗

OpenAI Introduces Computer Use Tool to Agents API

Developers can now build AI agents that navigate websites and interact with browser interfaces within an OpenAI-hosted environment.

The update provides a framework for automating web-based tasks like testing or data collection, while placing the responsibility for security limitations and action-scope control on the developers implementing the tool.

Full story ↗Original source ↗

OpenAI: Ultrafast is available for GPT-6 Astra and GPT-6.1 Sol, with low rate limits

OpenAI describes Ultrafast as its fastest API service tier. It is available to all API users for GPT-6 Astra and GPT-6.1 Sol at low rate limits.

Developers can select OpenAI’s fastest API service tier for both GPT-6 Astra and GPT-6.1 Sol, considering higher cost, low rate limits, and regional processing restrictions that differ by model.

Full story ↗Original source ↗

GPT-6.1 Sol API: 1,050,000-token context, pricing tiers and developer limits

OpenAI's developer docs list GPT-6.1 Sol's context window, output limit, per-token prices, long-prompt surcharge, Fast and Ultrafast multipliers, and per-tier rate limits.

Teams planning costs and capacity need to weigh several factors together: context length, service mode, region and usage tier each affect price or throughput. These are terms from OpenAI's documentation; the source material includes no independent evaluation of the model's performance.

Full story ↗Original source ↗

GPT-6.1 Sol Model Now Available on Amazon Bedrock

The GPT-6.1 Sol model has been added to Amazon Bedrock, providing developers with enhanced reasoning capabilities and lower task costs.

This release allows developers to execute complex multi-step business tasks with greater efficiency and reduced operational costs compared to previous versions.

Full story ↗Original source ↗

Grok 4.7 Model Added to Amazon Bedrock

Amazon Bedrock now supports Grok 4.7, featuring a 500K token context window and adjustable reasoning effort for developers and enterprise users.

This update provides professional users with granular control over AI reasoning intensity and extended context handling, which are essential for building robust agent-based applications.

Full story ↗Original source ↗

AWS Adds Claude Sonnet 5.5 to Amazon Bedrock

Amazon has integrated Claude Sonnet 5.5 into its Bedrock service, offering lower costs and faster processing for coding and documentation tasks.

This update provides enterprises with a more cost-effective way to automate complex technical workflows. By lowering operational overhead for coding and analysis, developers can integrate high-performance AI capabilities into their AWS workloads more sustainably.

Full story ↗Original source ↗

GitHub Copilot Adds Claude Sonnet 5.5 Model

Anthropic's Claude Sonnet 5.5 is now available in GitHub Copilot for eligible professional and enterprise users via usage-based billing.

This integration allows developers to leverage Claude Sonnet 5.5 directly within their coding environments, offering greater flexibility in model selection.

Full story ↗Original source ↗

NVIDIA Announces Open Agent Safety Platform

NVIDIA announced an open software platform and reference system design to govern AI agents from testing through deployment across software and infrastructure.

The platform adds governance and control beyond the model and agent harness, including at the infrastructure level. This gives organizations tools to define and enforce boundaries as autonomous agents work.

Full story ↗Original source ↗