Claude Opus 5: A Brilliant Yet Frustrating Model in Use
Claude Opus 5, the latest artificial intelligence model, has recently been put to the test. After a thorough evaluation, here’s an overview of the performance and features of this model.
A ceiling of intelligence reached?
One of the notable observations is the impression that a ceiling has been reached in terms of model intelligence. This saturation raises questions about the variables that truly matter now in the evaluation of these technologies.
A complex personality
Opus 5 stands out for a personality that some might describe as "neurotic." During coding sessions, the model exhibited hesitations, particularly refusing to resolve certain merge conflicts, which can be frustrating for users seeking smooth assistance.
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.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Comparison with other models
In a comparative test, Opus 5 was pitted against other models such as GPT-5.6 Sol, Sonnet 5, and Gemini 3.1 Pro. These models were evaluated within the framework of the HIA ranking, which is based on seven models and six tasks, with blind scoring. The results revealed varied performances, with Opus 5 only standing out in one use case where it achieved the highest score.
Interactions and verbosity
During an interaction, Opus 5 and GPT-5.6 Sol were both asked the same question, "who is smarter, you or me?", which highlighted certain limitations of these models. Additionally, the verbosity of Claude Slop, a version of Opus 5, proved irritating, underscoring a recurring issue in the use of this model.
A mixed verdict
The model's uneven performance raises questions about its widespread adoption.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.