AI News

Anthropic, OpenAI and Google expanded their workplace products and lowered prices, while a dispute over three fired OpenAI safety researchers kept attention on how labs handle outside oversight.

Top headlines

1. Anthropic launches Claude Dashboards and Claude Motion

   Anthropic says Claude Dashboards builds dashboards from company data in warehouses such as BigQuery, Snowflake, Databricks and Redshift, keeps them current and shows the query behind each number; it is in beta on paid plans. Claude Motion, in beta on Team and Enterprise, generates animated explainers as editable code and exports MP4. Anthropic also moved Docs, Slides and Design out of beta on all plans, including Free, and says more than 45 million docs, decks and designs have been made.

   Covered by: Superhuman, AlphaSignal, Evolving AI · Source

2. OpenAI brings GPT-6 with Intelligent UI to all ChatGPT users

   OpenAI is rolling out GPT-6 to every ChatGPT user, including the free tier, together with an Intelligent UI that can answer with interactive charts, diagrams, buttons and small tools inside the conversation. Several newsletters report that paid tiers use GPT-6 Sol while Free and Go users get GPT-6 Luna.

   Covered by: Superhuman, AlphaSignal, Evolving AI, AINews (Latent Space) · Source

3. Anthropic releases Claude Haiku 5.5

   Anthropic prices Claude Haiku 5.5 at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100k tokens, rising five-fold above that, and estimates it costs about 75% less to run than Haiku 4.5 on average. Anthropic reports 72.4% on its OSWorld 2.1 computer-use subset versus 48.9% for GPT-6 Luna, and says it is the first Haiku model with an adjustable effort setting. Anthropic also halved Sonnet 5.5 cache-read prices.

   Covered by: AlphaSignal, Evolving AI, AINews (Latent Space), Superhuman · Source

4. OpenAI opens the Decisions API in public beta

   OpenAI's Decisions API returns typed answers instead of free text: a probability that a condition holds, a pick from a fixed list, or a score against ordered levels. According to OpenAI's guide it runs only on gpt-6-luna for now, is about 10x faster than the Responses API, and charges $0.10 per million input tokens with no output-token charge. Several other vendors shipped comparable decision models in the same period.

   Covered by: AlphaSignal, AINews (Latent Space) · Source

5. Fired OpenAI safety researchers dispute misconduct claims

   Jasmine Wang, Tomek Korbak and Mikita Balesni published an open letter disputing their dismissals and warning of a chilling effect on safety work, TechCrunch reports. OpenAI told TechCrunch the three violated policies on handling sensitive company information and that an investigation found a pattern of misconduct. The researchers urged OpenAI to keep commitments on third-party auditors and model monitorability; an internal OpenAI memo says it agrees with those recommendations.

   Covered by: AINews (Latent Space), Superhuman, Evolving AI · Source

6. Anthropic splits its Cyber Verification Program into three access tiers

   Anthropic's expanded program offers Defense Access, Red Team Access for authorized penetration testing, and Specialized Access for critical-infrastructure testing reviewed with the US government, each with fewer cyber safeguards on Opus 5.5, Sonnet 5.5 and Mythos 5.1. Project Glasswing is merged into the program. Anthropic says Glasswing partners found at least 129,000 verified vulnerabilities between April and July 2026.

   Covered by: AlphaSignal · Source

7. Anthropic updates its Usage Policy, including a rule on abuse of Claude

   Anthropic's updated Usage Policy takes effect on November 12. It prohibits sustained and needless abusive or cruel behavior toward Claude in extreme cases, while ordinary frustration, pushback, dark fiction and model testing remain allowed; Claude's ability to end abusive conversations is the main enforcement tool. Other changes cover deceptive influence campaigns, weapons, surveillance and high-risk uses.

   Covered by: Evolving AI, Superhuman · Source

8. Google Cloud unveils a universal Gemini agent for work

   At Gemini at Work 2026, Google Cloud introduced the Gemini agent, which it describes as a single agent for knowledge work, content creation and coding that plans multi-step tasks, connects to business systems and returns finished work in the tools people already use. Google says it selects the best model for each task and includes cost controls and enterprise governance.

   Covered by: Superhuman, AINews (Latent Space) · Source

9. Anthropic commits $150M to the US Genesis Mission

   Anthropic says it will provide $150 million over three years in Claude, Claude Code and API credits to research projects across more than 15 federal agencies in the Genesis Mission, naming NASA, NIH and NSF, along with joint work on areas such as fusion energy and quantum computing.

   Covered by: Evolving AI, AINews (Latent Space) · Source

10. Biohub expands its virtual cell initiative to $1.8 billion

    Biohub says its initiative to generate AI-ready biological data and build predictive models such as a virtual cell now totals $1.8 billion in funding, data and compute. Google DeepMind, Isomorphic Labs and Meta are contributing $300 million collectively, and the US Department of Energy more than $500 million over five years.

    Covered by: Evolving AI · Source