Anthropic unveils Opus 5, a potential rival to the controversial Fable 5

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Anthropic Unveils Opus 5, a Potential Rival to the Controversial Fable 5
Following government concerns regarding cybersecurity, Anthropic has launched its latest model, Claude Opus 5, which boasts "near" capabilities to Claude Fable 5 in many areas while significantly outperforming it in complex coding tasks. Fable 5, the Mythos-class model that sparked government outrage, was taken offline for several weeks before being reintroduced with even more robust cybersecurity protections.
The concerns surrounding Fable 5 and weeks of negotiations between Anthropic and the government marked the beginning of a new era of AI regulation under the Trump administration. Shortly thereafter, OpenAI launched GPT-5.6, which underwent a trial period of about two weeks, being deployed in limited preview only for government-approved entities.
Anthropic Promises Enhanced Cybersecurity Protections with Opus 5
When asked if Anthropic had submitted Opus 5 to the Trump administration before its public launch, Anthropic spokesperson Danielle Ghiglieri told The Verge: "We continue to work with our government partners to conduct their own independent testing of our models. This includes Opus 5."
Anthropic is particularly targeting its enterprise clients with Opus 5, claiming it is ideal for knowledge work and biology, representing the perfect everyday model, while Fable 5 remains the best for the most ambitious tasks and long-term AI agent projects. The company specified in a statement that Opus 5 has stronger cybersecurity protections than its predecessor, Opus 4.8, and is "the most aligned Opus model and least likely to be deceived for misuse."
With the launch of this new model, the company also seems to subtly address recent controversies regarding supposedly hidden usage caps—which have led to at least one lawsuit—as well as financial pressures on AI labs, or concerns that the costs of AI labs are being passed on to their clients in new ways.
- The price of Opus 5 per million tokens is the same as its predecessor ($5 for input / $25 for output).
- Anthropic emphasized that it is "half the price" compared to Fable 5.
- It is also slightly cheaper than OpenAI's GPT-5.6.
Customers can also opt for the "Fast Mode" for Opus 5 (a feature launched today in research preview) to achieve higher speeds for double the price. There is also an option to automatically revert to a lower-tier model when protections require Opus 5 to deny a user request.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.