Anthropic Unveils Claude Opus 4.8 With Sharper Reasoning, Lower AI Costs
The model is available immediately across Claude.ai and the Claude API.
Anthropic has unveiled Claude Opus 4.8, the latest version of its flagship AI model, introducing improvements in coding, reasoning, agentic workflows and AI safety while keeping pricing unchanged from its predecessor.
The company said Opus 4.8 builds on Claude Opus 4.7 with stronger performance across benchmarks and improved reliability when handling long-running and autonomous tasks. The model is available immediately across Claude.ai and the Claude API.
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.
— Claude (@claudeai) May 28, 2026
Available today at the same price. pic.twitter.com/EufxL7T1kb
Alongside the model launch, Anthropic introduced several new features, including “dynamic workflows” in Claude Code, which allows the AI system to plan and execute large-scale tasks using hundreds of parallel subagents within a single session.
According to the company, the feature can handle codebase-scale migrations involving hundreds of thousands of lines of code while verifying outputs against existing test suites. Dynamic workflows are currently available in research preview for Enterprise, Team and Max users.
Been building with Claude Opus 4.8 (out today!) for a couple weeks and it's already the model I reach for first. The best part is how much I can just let it run. It's more honest about its own work, flags what it's unsure of, & catches flaws in its code before handing it back.
— Mike Krieger (@mikeyk) May 28, 2026
Anthropic also added a new “effort control” setting on Claude.ai and Cowork, allowing users to choose how much computational effort Claude applies to a task. Higher effort settings improve reasoning depth and output quality, while lower settings prioritise faster responses and lower token usage.
The company said one of the most significant improvements in Opus 4.8 is model honesty and reliability. Internal evaluations showed the model is around four times less likely than Opus 4.7 to overlook flaws in code it generates or make unsupported claims during tasks.
Anthropic’s Alignment team said the model “reaches new highs on our measures of prosocial traits like supporting user autonomy and acting in the user’s best interest.”
The launch also includes API updates that allow developers to modify Claude’s system instructions mid-task without interrupting prompt caching or workflow execution.
Anthropic said it is also working toward releasing “Mythos-class” AI systems with significantly higher intelligence levels in the coming weeks.
Recently, the startup also raised $65 billion in Series H funding led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, valuing the company at $965 billion post-money.
