What happened
Anthropic released Claude Opus 4.8 on May 28, 2026. It builds on Opus 4.7 and improves across benchmarks while keeping the same price. Anthropic describes it as a better collaborator with sharper judgment, more honesty about its own progress, and the ability to work alone for longer.
Benchmark gains include agentic coding rising from 64.3 percent to 69.2 percent, reasoning with tools moving from 54.7 percent to 57.9 percent, and computer use moving from 82.8 percent to 83.4 percent. A fast mode lets the model run at 2.5 times the speed, now three times cheaper than before.
Why it matters
Small gains at the top of the model range still matter. Many teams run Opus for hard tasks like coding, research, and long agent runs. A model that is more honest about uncertainty is safer for high stakes work, since it is less likely to state wrong things with confidence.
Cheaper fast mode also lowers the cost of running Opus on big jobs.
MintedBrain take
For heavy Claude users, Opus 4.8 is a clear upgrade with no price change. The honesty gains are the most useful part. A model that says when it is unsure saves time and reduces risk. Test it on your hardest tasks to see the difference.
Discussion
Sign in to comment. Your account must be at least 1 day old.