Your daily update on the latest from Anthropic’s Claude AI. The AI scene has been busy the last few weeks, with some big news on new model releases and some important regulatory moves.
CLAUDE OPUS 4.8 IMPROVEMENTS
On May 28, 2026, Anthropic announced a major upgrade to its Opus 4.7 model, called the Claude Opus 4.8. The new release is better across benchmarks, making it a more effective collaborator. Opus 4.8 Improves coding, agentic skills, reasoning, and practical knowledge work tasks. New additions include a user control of how hard Claude tries and “dynamic workflows” for Claude Code, which can execute hundreds of parallel subagents to solve very large-scale problems. Opus 4.8, the fast mode is now three times cheaper than previous models, and runs at 2.5 times the speed.
New Slack Integration: Mention Claude
Anthropic and Salesforce announced the launch of Claude Tag, an innovative integration for teams to work with Claude in Slack, on June 24, 2026. Users can call on Claude as a team member in any channel by typing @Claude, enabling real-time collaboration on workflows and ideation. This feature replaces the existing Claude in Slack app and it is described as a “ambient” colleague that can proactively offer updates when ambient behavior is switched on.
STATUS OF FABLE 5 AND MYTHOS 5
One of the big developments was Mythos 5 and Claude Fable 5. These advanced models were released on June 9 2026, but were quickly suspended on June 12 2026 by the US government. They were concerned about their automated hacking capabilities and that was the order. While a non-public version of Mythos 5 has since opened its access to some trusted partners for government agencies, the public availability of Fable 5 remained unclear. However, recent reports indicate that as of June 27, 2026, Claude Fable 5 is also expected to be back online soon, perhaps within the next week. This comes after positive collaboration between Anthropic and government officials to address initial safety concerns.
