Claude Opus 4.8 arrives with more honesty and coding power

Anthropic released Claude Opus 4.8 on May 28, 2026, the new version of its most advanced AI model. The update brings gains in coding, reasoning and — perhaps most interesting — honesty. And the best part: at the same price as the previous model.

Quick answer: what changed?

Opus 4.8 is more accurate in code, better at admitting when it is uncertain, and works independently for longer on complex tasks. It is an evolution of Opus 4.7, with higher benchmarks and new Claude Code features.

More honesty (really)

A highlight is the more trustworthy behavior. According to Anthropic, Opus 4.8 has sharper judgment and is more honest about its own progress. In practice, it flags uncertainties more often and makes fewer unsupported claims. In code, it is about four times less likely to let flaws pass unremarked — a huge gain for those who use AI to program.

The benchmarks

The numbers back the talk. Here is the comparison with Opus 4.7:

Benchmark	Opus 4.7	Opus 4.8
Agentic coding	64.3%	69.2%
Reasoning with tools	54.7%	57.9%
Agentic computer use	82.8%	83.4%
Knowledge work	1753	1890

New features

Beyond the model, useful functions arrived. On claude.ai, you can now control the effort level Claude puts into a task. Claude Code gained “dynamic workflows”, built for large-scale problems. And the Opus 4.8 fast mode runs at 2.5× the speed and is three times cheaper than in previous models.

Evolution, not revolution

Keep expectations balanced: experts described 4.8 as “a modest but tangible improvement”. It is not an everything-changes leap, but a solid refinement — exactly what to expect from an industry now shipping models every few weeks.

Why this matters to you

If you use AI for work, more honesty and fewer silent errors mean less rework and more trust. And the cheaper fast mode makes heavy daily use more viable.

Frequently asked questions

Is Claude Opus 4.8 more expensive?

No. Anthropic launched it at the same price as Opus 4.7, with better performance.

What improved in coding?

It is about four times less likely to let flaws pass unremarked, and scores higher in agentic coding.

What is fast mode?

A mode that runs at 2.5× the speed and is three times cheaper than in previous versions.

At DigitalRadar, we follow the AI race closely. Stay on the radar so you do not miss the next updates.