Fable 5: Automation Record, Humans Remain Key

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
Fable 5: A New Milestone in Freelance Work Automation
Anthropic recently unveiled the impressive performance of its Fable 5 model, which has set a new record in automation within the freelance work sector. This model achieved an automation rate of 16.1%, marking a significant advancement over previous AI models.
Remarkable Performance of Fable 5
Fable 5 has managed to double the automation rate of its predecessor, Opus 4.8, which reached 8.3%. In comparison, GPT-5.5, another competing model, scored 6.3%, placing it in third position. These results reflect Fable 5's enhanced ability to autonomously execute complex tasks.
Testing Framework
The Center for AI Safety (CAIS) evaluated Fable 5 through its Remote Labor Index (RLI), a tool published in October 2025. This index assesses the ability of AI agents to complete real and economically viable freelance projects. The areas covered include computer-aided design, graphic design, data analysis, and video production. The deliverables generated by the models are then judged by human experts according to strict professional standards, and the automation rate reflects the proportion of projects where the AI matched or surpassed human work.
Test Results and Implications
The tests involved Fable 5, GPT-5.5, and Opus 4.8 in various tasks such as creating a 3D model of an engagement ring, producing a video advertisement, and mapping a floor plan. Even though Fable 5 experienced a government-mandated interruption in mid-June, its partial results demonstrated its superiority over other models.
Impact on the Freelance Market
While Fable 5 has achieved an impressive automation rate, this does not mean that freelance jobs are at risk of extinction. With a rate of 16%, there is still a long way to go before achieving full automation. Moreover, the integration of AI tools within organizations is often hindered by security concerns and other technical obstacles.
For AI to fully replace human freelancers, companies would need to establish a network of agents capable of verifying work quality, adhering to budgets, and managing deadlines.
Current Limitations and Future Perspectives
The CAIS attempted to replace human evaluators with a "judge LLM," but this effort ended in failure. The evaluation of RLI deliverables requires advanced computing skills that current models do not yet fully master. However, with ongoing investments in more capable models, these skills could improve rapidly, thereby reducing certain freelance opportunities.
Conclusion: An Uncertain Future
It is difficult to predict the future of freelance work in the face of AI. Some tasks that are quick for a qualified professional remain out of reach for current AI, while others, more time-consuming for humans, can be executed in record time by these models. The evolution of AI model capabilities could transform the freelance work landscape, but for now, human input remains indispensable.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.