---
title: "Introducing Claude Opus 5"
description: "Claude Opus 5 is available as a new model positioned near the frontier intelligence of Claude Fable 5, at half its price, while remaining behind Mythos 5 on cybersecurity tasks. Reported results show state-of-the-art performance on coding and knowledge-work evaluations, including Frontier-Bench, alongside strong results on ARC-AGI 3, Zapier AutomationBench, and OSWorld 2.0. The model uses configurable effort settings to trade intelligence for token use, speed, and cost, and is described as stronger at verifying work, iterating, and handling long agentic tasks. It costs $5 per million input tokens and $25 per million output tokens, with Fast mode running about 2.5 times faster at twice the base price. Safeguards allow vulnerability finding while blocking binary-based scanning, penetration testing, and exploit generation, with flagged requests able to fall back to Opus 4.8."
---

# Introducing Claude Opus 5

[Anthropic](https://yomu.fyi/company/anthropic) · Jul 25, 2026

**Type:** Announcement

## Summary

Claude Opus 5 is available as a new model positioned near the frontier intelligence of Claude Fable 5, at half its price, while remaining behind Mythos 5 on cybersecurity tasks. Reported results show state-of-the-art performance on coding and knowledge-work evaluations, including Frontier-Bench, alongside strong results on ARC-AGI 3, Zapier AutomationBench, and OSWorld 2.0. The model uses configurable effort settings to trade intelligence for token use, speed, and cost, and is described as stronger at verifying work, iterating, and handling long agentic tasks. It costs $5 per million input tokens and $25 per million output tokens, with Fast mode running about 2.5 times faster at twice the base price. Safeguards allow vulnerability finding while blocking binary-based scanning, penetration testing, and exploit generation, with flagged requests able to fall back to Opus 4.8.

## Context

Claude Opus 5 is being introduced as an everyday model that improves performance and cost-effectiveness over Opus 4.8, while supporting coding, knowledge work, scientific research, and agentic tasks. The announcement also addresses cybersecurity and biology safeguards, including limitations on exploit development and long-running autonomous research.

## Approach / What changed

Claude Opus 5 uses configurable effort settings to balance intelligence, token consumption, speed, and cost. It is offered across platforms and through the Claude API, with Fast mode, automatic safety fallbacks, and mid-conversation tool changes available as part of the launch. Its safeguards permit source-code vulnerability discovery but restrict binary-based scanning, penetration testing, and exploit generation.

## Takeaways

- Effort settings let customers optimize for intelligence or conserve tokens for faster, cheaper results; on CursorBench 3.2 at maximum effort, Opus 5 came within 0.5% of Fable 5’s peak score at half the cost per task.
- On life-sciences evaluations, Opus 5 outperformed Opus 4.8 across the set, including gains of 10.2 percentage points on spectroscopy-based organic chemistry tasks and 7.7 points on protein-variation tasks.
- Developers can use claude-opus-5 through the Claude API; beta updates add mid-conversation tool changes without invalidating the prompt cache and automatic fallback routing for classifier-flagged requests.

**Tags:** [AI](https://yomu.fyi/topic/ai), [API](https://yomu.fyi/topic/api), [LLMs](https://yomu.fyi/topic/llm), [Model Evaluation](https://yomu.fyi/topic/model-evaluation), [Security](https://yomu.fyi/topic/security)

- Source: [Anthropic](https://www.anthropic.com/news/claude-opus-5)
- Source URL: https://www.anthropic.com/news/claude-opus-5
- Ingested by Yomu: 2026-09-09T11:36:54.627Z

[Read original post](https://www.anthropic.com/news/claude-opus-5)