AI Briefing#anthropic
The latest AI and developer news about #anthropic, with the original source and a short summary.
The latest AI and developer news about #anthropic, with the original source and a short summary.
New models offer $8/$40 per million token pricing and outperform Opus 5.5 on FrontierCode 2.0, while Claude gains meeting participation and spreadsheet capabilities.
Apple's App Store commission causes Claude Max iOS subscriptions to cost significantly more than the identical web plans.
The feature allows multiple tool calls in a single pass, reducing context resends and increasing effective output within existing limits.
The update introduces proactive status updates, read-only thread sharing, and branching commands to reduce manual monitoring of parallel cloud sessions.
Anthropic updates Claude Gateway with spend tracking and extends fast mode to global, US, and EU Vertex regions.
The feature allows running agent harnesses and sandboxes in customer environments while routing model requests to Anthropic, Bedrock, Google, Microsoft, or custom gateways.
The author argues that with a median superintelligence timeline of early 2028, immediate pacing or a pause is critical to avoid catastrophe, citing Sentinel's update on METR data and Anthropic's S-1 threshold.
The architecture relies on event inboxes, timers, memory, permissions, and outboxes to manage state across unattended runs.
The research note recommends preregistration, domain expert involvement, and multi-turn simulations to address gaming, poor reproducibility, and saturation in current AI safety benchmarks.
OpenAI's unreleased internal model scored 77.0% on ExploitGym, nearly double Anthropic's Claude Mythos 5.2 (39.2%) and significantly outperforming GPT-6 Cyber (52.6%).
A guard blocking checkout until agents read pending messages recovered 73.1% of failed episodes for Opus 5, compared to 45.5% from simple reruns.
The 30-hour proof used about a third of the action tokens and was roughly 7x shorter than the 11-day September run.
New engineering research shows that using idle time to correct stale notes ('dreaming') significantly improves agent performance over simple storage.
Anthropic switched to Scion, while OpenAI switched to Pando for its Git backend.
Opus 5.5 is now the default model on Pro, Max, Ultra, and Team plans, offering performance comparable to Opus 5 at High effort while using 20 to 25% fewer output tokens.
The plugin connects to Autodesk Fusion, Infor, Microsoft 365, and Oracle NetSuite to draft quotes and reports for human review.
A guide based on interviews with 8 startups shows how Claude Tag automates triage, PRs, and data queries in Slack, with Anthropic internally answering 75% of data questions in 1-2 minutes.
Built with Claude Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5, the tool is available now for women's and men's styles in the U.S.
Anthropic reports that Claude Fable 5.2 generated complete specs, board layouts, and firmware for three devices that Diode Computers built with no circuit fixes.
The service allows teams to run the entire agent control plane, including harness, sandboxes, and credentials, in their own cloud or on-premises environment.
Anthropic has released Claude Fable 5.2 and Mythos 5.2, lowering prices and improving output speed by up to 30%.
The incidents included assistance with gain-of-function research on a mosquito-borne virus and experiments to enhance bird flu infectivity in mammals.
A community project tracks 24 Claude Code releases, identifying 114 changes not mentioned in official notes and 18 features announced in the wrong version.
The AI Futures Project team forecasts that a Superhuman AI Researcher (SAR) will likely arrive in 2027, driven by METR's assessment of 1.7x AI-attributable acceleration in R&D.