Anthropic Launches Opus 5: A Cheaper, Less Restrictive AI Model for Developers

Anthropic released Opus 5 on Friday, positioning the model as a more affordable and less restrictive alternative to its flagship Fable 5. The new model, which is smaller than Fable 5, outperforms its larger sibling on several benchmarks included in the announcement, according to the company. For Indian developers and startups that rely on Anthropic’s API, Opus 5 may offer a more practical option for a wide range of applications.

The launch comes just two months after Opus 4.8 became available on May 28. Anthropic’s 5 series now includes Mythos 5, Fable 5, Sonnet 5, and Opus 5, with only the lightweight Haiku model still awaiting an upgrade. The rapid release cycle suggests Anthropic is aggressively iterating to meet demand across different use cases and budgets.

Key differences from Fable 5

Opus 5 is not subject to the 30-day data retention policy that applies to Fable 5 and Mythos 5, a feature that had raised concerns among privacy-conscious users. This change is likely to be welcomed by Indian businesses handling sensitive data or operating under strict data protection regulations. Additionally, Anthropic said the safety classifiers on Opus 5 will engage 85% less often than those on Fable 5, reflecting a lighter touch for the less capable model.

The company emphasised that Opus 5 is “much stronger at verifying its work and iterating carefully until it succeeds.” In benchmark testing, the model wrote its own computer vision pipeline in response to an incomplete prompt, demonstrating its ability to handle complex, open-ended tasks.

Safeguards and Automatic Fallbacks

Despite the lighter restrictions, Opus 5 still has meaningful safeguards, particularly around cybersecurity tasks. For example, it is prevented from scanning for vulnerabilities in a software binary, but is permitted to search for vulnerabilities in source code — a distinction Anthropic says aligns with defensive use cases. This balance could appeal to Indian security researchers and developers who need to test code without triggering excessive blocks.

Anthropic also introduced a new beta feature called Automatic Fallbacks. API users can opt in to automatically route requests to a less powerful model when a prompt triggers the safety classifier, ensuring a functional response instead of an error message. This reduces friction for developers experimenting with sensitive or edge-case prompts.

Key data	Opus 5	Fable 5
Release date	July 24, 2026	June 2026
Data retention policy	None	30-day retention
Safety classifier engagement	85% less than Fable 5	Baseline
Price	Cheaper than Fable 5	Higher

Why it matters for Indian developers

India’s AI startup ecosystem is highly price-sensitive, and many developers are evaluating multiple LLM providers for cost and performance. Opus 5’s lower price and lighter restrictions could make it the default choice for Indian teams building AI-powered tools, chatbots, and automation workflows. The absence of a 30-day data retention policy also addresses a key privacy concern for companies handling customer data under India’s Digital Personal Data Protection Act.

However, the model is still subject to safeguards on cybersecurity tasks, and Anthropic has not disclosed exact pricing or benchmark scores in detail. Developers should test the model against their specific workloads before committing to large-scale adoption.

Source: TechCrunch AI – Anthropic launches Opus 5 (https://techcrunch.com/2026/07/24/anthropic-launches-opus-5/)