TechCrunch · AIJul 25, 2026, 01:00 UTC+8

Anthropic launches Opus 5, cheaper and less restrictive than Fable 5

Anthropic launches Opus 5

Anthropic released Opus 5 on July 24, just two months after Opus 4.8. It beats Fable 5 on several benchmarks, is not subject to the 30-day data retention policy, and its safety classifiers are expected to trigger 85% less often. Cheaper and less restrictive, it will be the better pick for most use cases. A beta 'Automatic Fallbacks' feature also routes blocked prompts to a weaker model.

Why it matters: Anthropic flagship model refresh: smaller than Fable 5 yet beats it on benchmarks, cheaper, with looser data policy. TechCrunch exclusive, cross-source cluster will follow. Missing exact pricing and benchmark figures keeps it from 90+, but still a same-day must-write.

Read the original ↗Export Markdown

Take

The reason to click: Anthropic is drawing a clear line between its own models. Opus 5 is smaller than Fable 5 but scores higher on the company's published benchmarks, costs less, and drops two big friction points — no 30-day data retention policy, and safety classifiers that trigger 85% less often. The message is pretty blunt: for most work, use Opus 5 and stop fighting Fable's restrictions.

I'd discount the benchmark claims a bit for now. These are Anthropic's own numbers, no third-party evals yet, so we don't know how real-world reasoning holds up against Fable 5. The article also doesn't give actual pricing — just "cheaper" — so it's unclear whether that's a meaningful cut or a token gesture.

The beta Automatic Fallbacks feature is a pragmatic touch: when a prompt gets blocked by the safety classifier, it routes to a weaker model instead of refusing outright. Smart idea, but the article doesn't say whether the fallback responses are actually usable.

Source: TechCrunch · AI · techcrunch.com
Published: Jul 25, 2026, 01:00 UTC+8 · Jul 25, 2026