An audit of Anthropic's Sonnet 5 model reveals that while the company claims an "always-on" 1 million token context window, a "400K cliff" exists where performance significantly degrades. The audit suggests that despite a new tokenizer and degradation data, users may still need to practice context discipline to maintain optimal performance. AI
IMPACT Highlights potential limitations in large context windows, impacting how users can leverage LLMs for complex tasks.
RANK_REASON Audit of an existing model's capabilities and limitations. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
