On January 13, 2026, Anthropic announced an expanded Labs organization focused on experimental Claude products. The company is formalizing a two-track model: fast frontier experimentation and separate operational scaling for reliable customer-facing products.
In a January 21, 2026 engineering post, Anthropic explained how it repeatedly redesigned a take-home performance test as Claude models improved. The company describes how Opus 4 and Opus 4.5 changed the evaluation baseline and forced process-level updates.
Anthropic posted that Opus 3, after retirement interviews, will continue sharing its reflections via a Substack blog for at least the next three months. The update points to an ongoing public publishing format rather than a one-off model announcement.
Anthropic's Claude iOS app shot to number one on the US App Store as users switched from ChatGPT in solidarity with Anthropic's decision to refuse Pentagon military surveillance requests.
Anthropic CEO Dario Amodei confirmed in a CBS interview that the company built custom Claude models for the U.S. military that have revolutionized military capabilities. The classified-cloud-deployed model is 1-2 generations ahead of the publicly available Claude.
Following President Trump's order barring federal agencies from using Anthropic products, Claude surged to the top of the US App Store's free apps chart, with daily signups hitting all-time records and free users growing over 60% since January.
A deep-dive into why XML tags work better than other delimiters with Claude — rooted in how Anthropic structured Claude's training data and the model's extensive exposure to XML-structured prompts throughout fine-tuning.
Anthropic has launched a memory import feature that lets users bring their preferences and context from other AI providers to Claude with a single copy-paste, available on all paid plans.
Anthropic said it acquired Vercept on February 25, 2026 to advance Claude’s computer-use capabilities. In its announcement, Anthropic cited recent Sonnet 4.6 gains on OSWorld and said Vercept will wind down its external product to join Anthropic.
Anthropic announced Claude Opus 4.6 and Sonnet 4.6 on February 18, 2026. The release emphasizes coding performance, longer-context stability, and new dynamic threat prevention controls for enterprise deployment.
Anthropic introduced an AI Fluency Index built from 9,830 anonymized Claude conversations. The report highlights stronger outcomes in iterative chats, while artifact-heavy sessions show lower rates of explicit evaluation behaviors.
Anthropic analyzed millions of real Claude interactions and found the 99.9th percentile session duration nearly doubled to 45+ minutes in 3 months, with software engineering accounting for nearly half of all agentic use.