AI Product Updates: New Features and Releases

Claude Opus 5.5 matches Anthropic's Fable 5.1 on most work tasks and costs 40% less per task than Opus 5, priced at $4 and $20 per million input and output tokens.

Autosana's testing product now lets users run tests on two devices at once in real time.

Claude Opus 5.5 is now the Standard effort level in Perplexity Computer, scoring 0.610 on the WANDR benchmark at $4.13 per task, 67.6% cheaper than Claude Fable 5.1.

DigitalOcean Managed Agents is in public preview, running Claude Code, Codex, or LangGraph in isolated environments that pause when idle, with credentials kept away from the agent.

Musk announced that Grok's Bot feature is now available inside Tesla cars.

Vercel's v0 generative frontend tool now offers Claude Opus 5.5 as a selectable model that users can switch to directly.

Anthropic released Opus 5.5, now available in Claude Code and the Claude Platform, pitched as more intelligent, lower priced, and faster.

On the Claude Platform, Opus 5.5 is now $4/$20 per million tokens, with cache reads down 60% from $0.50 to $0.20; run /claude-api migrate to switch.

Thinking is always on for Opus 5.5 with medium as the default and an adjustable effort setting; long-running agents use compaction, and fast mode runs up to 2.5x at $8/$40.

Opus 5.5 is the first Opus model with Fable 5.1-class safeguards on cyber, bio, and frontier LLM development, and flagged requests fall back to another model.

Anthropic released Opus 5.5, which performs at the level of Fable 5.1 and is about 30% faster and 40% cheaper per task, while Claude Code 5-hour limits rise 20% and Pro, Max, and Team users get a reset.

Pro, Max and Team users can claim a reset in Settings → Usage until Oct 22, and Opus 5.5 is now the default paid-plan model, stretching the 5-hour and weekly limits by 25%.