Claude’s post announces Fable 5 as state-of-the-art on most benchmarks, showing strong results in software engineering, knowledge work, scientific research, and vision.
Performance gains over previous models widen significantly as tasks become longer and more complex, according to the announcement.
The included benchmark table positions Mythos 5/Fable 5 ahead of Opus 4.8, GPT 5.5, and Gemini 3.1 Pro in agentic coding, multidisciplinary reasoning, and cybersecurity-related metrics.
