Anthropic launches Claude Opus 5, a cheaper AI model for coding, agents and enterprise workflows

2026-07-24 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

Anthropic launched Claude Opus 5 on Friday, July 24, 2026, positioning it as a highly intelligent yet cost-effective AI model. Priced at \$5 per million input tokens and \$25 per million output tokens, matching Opus 4.8, Opus 5 offers near-Fable 5 intelligence at half the cost. It is now the default on Claude Max and the strongest model on Claude Pro. Benchmarks show Opus 5 achieving 43.3% on Frontier-Bench v0.1, more than doubling Opus 4.8's 18.7% and surpassing Fable 5's 33.7%, while also scoring three times higher on ARC-AGI 3. The model emphasizes efficiency, with customers reporting significant token savings and improved accuracy. Opus 5 also features self-verifying agent capabilities, reducing human review costs. Anthropic highlights its safety strategy, including intentional capability gaps and a 2.3 score on misaligned behavior, with fallbacks to Opus 4.8 for risky requests. This launch reflects Anthropic's focus on economical, daily-use AI for enterprise workflows.

Key takeaway

For AI Engineers and Directors evaluating large language models for enterprise deployment, Claude Opus 5 presents a strong economic argument. You should prioritize Opus 5 for most complex daily tasks and agentic workflows. Its near-Fable 5 intelligence at half the cost significantly reduces inference expenses. Self-verifying capabilities also minimize human review, accelerating automation. Reserve Fable 5 for only the most ambitious, long-horizon autonomous projects that outrun standard benchmarks.

Key insights

The AI market prioritizes economical, near-frontier intelligence for daily enterprise workflows over peak, expensive capabilities.

Principles

• Cost-efficiency and sustained coherence define model utility for enterprise.
• AI safety integrates intentional capability gaps and model fallbacks.
• Benchmarks excel at bounded tasks; long-horizon jobs demand sustained autonomy.

In practice

• Deploy Claude Opus 5 for complex, bounded daily tasks.
• Reserve Claude Fable 5 for multi-day, long-horizon autonomous projects.
• Adjust Opus 5's "effort" setting to optimize for speed and token savings.

Topics

• Claude Opus 5
• Enterprise AI
• AI Agents
• Inference Costs
• AI Safety
• LLM Benchmarking

Code references

• xlang-ai/OSWorld-V2

Best for: CTO, Machine Learning Engineer, VP of Engineering/Data, Director of AI/ML, AI Engineer, Consultant

Related on AIssential

• A Cheaper Claude, a Cracked Agent, and $52 Million for Your Iris — Frontier AI models are achieving higher performance at significantly reduced costs, making advanced capabilities more accessible.
• Anthropic's Opus 5 surprise — Claude Opus 5 delivers Fable 5-tier intelligence at half the cost, setting new benchmarks in agentic tasks and complex problem-solving.
• Introducing Claude Opus 5 — Claude Opus 5 delivers near-frontier intelligence with superior cost-efficiency and enhanced agency across diverse professional tasks.
• Anthropic Shipped Four Models in Nine Weeks. Opus 5 Explains the Other Three. — Claude Opus 5 offers near-frontier intelligence at half the price of Fable 5, impacting the value proposition of other models.
• Anthropic's Claude Opus 4.8 is here with 3X cheaper fast mode and near-Mythos level alignment — Claude Opus 4.8 offers a 3X cheaper fast mode and new subagent capabilities, alongside modest performance and significant alignment improvements.
• Introducing Claude Opus 5 on AWS: Anthropic’s most capable Opus model — Claude Opus 5 offers frontier AI capabilities for complex enterprise tasks, particularly in coding and long-running agentic workflows, with robust data governance on AWS.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.