# Claude Opus 5: A New Milestone in AI

Claude Opus 5 stands out with its performance and cost efficiency.

- **Source:** [Hacker News](https://news.ycombinator.com)
- **Published:** 2026-07-24
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://www.anthropic.com/news/claude-opus-5>
- **On Brifing:** https://brifing.si/s/9aebfded-6b74-4ca1-a356-731ecb4b1959

## Launch of Claude Opus 5

Claude Opus 5 is launched as a model that approaches the frontier intelligence of Claude Fable 5 at half the price. It sets a new standard in coding and knowledge work evaluations.

## Performance and Cost Efficiency

Opus 5 provides a significant performance boost at the same cost as its predecessor, Opus 4.8. Charts showing performance changes based on the model's effort setting allow users to optimize for intelligence or token savings.

## Superiority in Software Engineering Tasks

On Frontier-Bench v0.1, Opus 5 surpasses all other models and doubles the performance of Opus 4.8. On CursorBench 3.2, it comes within 0.5% of Fable 5's peak score.

## Knowledge Work and Problem Solving

On ARC-AGI 3, Opus 5 scores three times higher than the next best model in solving novel problems. On Zapier AutomationBench, it completes 1.5 times more tasks at the same cost.

## Advancements in Scientific Research

Opus 5 shows better performance than Opus 4.8 in life sciences evaluations, particularly in organic chemistry and protein-related tasks, with notable improvements.

## Strong Visual Outputs

Opus 5 makes significant progress in producing visual outputs, delivering stronger and more impressive results compared to previous models.

## Work Verification and Iteration

Opus 5 is much stronger at verifying its work and iterating carefully. This feature stands out in evaluations and early access testing.

## Real Bug Fixing

Opus 5 identifies and fixes a real bug in a popular open-source package manager, addressing an edge case missed by the community's patch. Competing models only fix surface symptoms.

## Market Data Feed Creation

An engineer at a trading firm uses Opus 5 to create a market data feed for a new exchange in a single session. Previous models couldn't complete this task, but Opus 5 even builds its own test harness.

## Early Access User Experiences

Early access users report that Opus 5 approaches Fable-level performance on FrontierCode 1.1 and shows particular strength in difficult debugging tasks within Devin.

## Zapier AutomationBench Leadership

Opus 5 tops Zapier's AutomationBench leaderboard without spending more tokens than prior Claude models. It successfully completes a full churn-prevention sequence.

## Scientific Approach in Genomic Analysis

In genomic analysis work, Opus 5 behaves like a careful scientist, reaching for the right statistical tests and staying on track through long multi-step analyses.

## Sources covering this story

- https://www.anthropic.com/news/claude-opus-5 — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._