Anthropic Launches Fable 5.1: Improved Programming Capabilities and Lower Costs

Deep Tide TechFlow reports that on September 2, Anthropic launched the new version of its AI model, Fable 5.1, claiming better performance in programming and scientific tasks while being more cost-effective. In programming, Fable 5.1 excels at handling long-term, complex tasks, including software project development, code review, and intricate scenarios involving entire application codebases. Its scientific capabilities have also been enhanced, covering experimental design, simulation of experimental results, and the interpretation of complex charts and data tables. On the cost side, token pricing for enterprise users remains the same as Fable 5, but the cost of reprocessing previously handled information has been reduced by 75%. This aspect can account for more than half of token usage in long texts and complex tasks, and the price reduction is expected to significantly lower overall operational costs. In terms of safety, Fable 5.1 has upgraded its classifier, achieving more accurate identification of risky instructions with fewer false positives, enabling more efficient use for users in fields like cybersecurity and biology. Additionally, Anthropic plans to allow enterprise clients to store data from its most advanced models on their own cloud infrastructure rather than Anthropic's servers, while continuing to enforce safety checks.