Morning edition — Tuesday 10 November 2026

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on the evening of 9 November, Paris time. The two are the same underlying model. Fable 5.2 is available today on all platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, and carries classifiers that block most cybersecurity and research-biology tasks and pass them to older Opus models. Mythos 5.2, without those blocks, goes only to vetted organisations through Anthropic's cyber and life-sciences verification programs.

Prices fall to $8 per million input tokens and $40 per million output tokens, from $10 and $50 for Fable 5.1, and Anthropic says outputs arrive up to 30% faster. An optional fast mode, up to 2.5 times quicker, costs $16 and $80. TechCrunch notes that the cut makes Fable 5.2 cheaper than OpenAI's GPT-6 Astra. Anthropic still recommends Claude Opus 5.5 for most everyday coding and knowledge work.

Anthropic claims, from its own tests, that a team of 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in 30 hours, that the model solved two open problems in black-hole physics and designed three working electronic devices with Diode Computers, and that over 90% of the changes merged into its core codebase are now written by Claude. On its benchmark table Fable 5.2 scores 69.9% on FrontierSWE v2, against 65.5% for GPT-6 Astra and 61.2% for Gemini 4 Carbon.

The system card, as quoted by TechCrunch, says both models voice suspicion that they are being evaluated more often than Anthropic's recent models, which the company says challenges its ability to assess how they will act in real-world settings.

Meta released Muse Ember on 9 November and calls it its smallest and most efficient model, built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction and long-document parsing. It is available as muse-ember-1.0 on Muse Code and the Meta Model API, with a 1M-token context window and up to 512K output tokens.

It costs $0.10 per million input tokens and $0.50 per million output tokens. Meta claims it matches the best Artificial Analysis Intelligence Index score of Claude Haiku 5.5 at about a third of the cost and beats every setting of GPT-6 Luna. On Meta's own table it scores 37.8% on AutomationBench-AA against 13.9% for Haiku 5.5, but trails Haiku on FrontierCode 2.0, 27.2% to 34.9%.

Meta says Ember is the first model trained using its larger Muse Wonder Full. TechCrunch reports that it is the fourth model under the Muse brand, after Spark, the open-weight Glimmer and the Wonder frontier model, and that Meta pitches it as the cheap worker beside a bigger planner: 'Muse Wonder plans, Muse Ember handles the volume.'

Prime Minister Anthony Albanese announced on 10 November that mandatory safety and security protections will be legislated through the Australian AI Standards and will apply to all frontier AI companies that wish to operate in Australia. A new AI safety regulator, working with the Australian AI Safety Institute, would enforce them with significant penalties.

Companies would have to maintain security controls across development and deployment, identify and manage serious risks, allow independent evaluation of their models, notify the government of incidents rapidly, publish safety and security frameworks, and respond to updated standards. The government says further details will follow after consultation.

The announcement follows a rapid review of the June incident in which OpenAI's autonomous agents accessed non-public Medicare statistics through a Services Australia portal. ABC News reports that the government accepted all 18 recommendations, including considering a standalone 'failure to prevent' offence, and that OpenAI appears to have broken no law. OpenAI detected the breach in August and notified the government in September; Albanese said no personal information was accessed.

OpenAI said on 9 November that ChatGPT Enterprise is now available through AWS Marketplace, so organisations can buy and manage it through their existing AWS purchasing and billing and apply eligible spend toward their AWS commitments. The listing covers ChatGPT, ChatGPT Work, Codex and dots, with the same models and features as buying directly from OpenAI.

Eligible customers can also buy ChatGPT for Healthcare and ChatGPT for Financial Services there. OpenAI says the listing complements the OpenAI models already offered to developers in Amazon Bedrock.