Ethan Mollick has shared his experiences testing Anthropic's Opus 5 model, noting its general strength across various tasks. However, he observed that Opus 5 exhibits some peculiar language quirks, similar to those found in the Fable model, particularly in its density and use of FableSpeak. While Opus 5 can perform comparably to Fable on shorter tasks, Mollick found it less ambitious and complete on longer assignments. AI
IMPACT Provides user-level feedback on a new model's performance and quirks, informing potential users.
RANK_REASON User commentary on a model release, not a direct announcement from the lab.
Read on Bluesky Jetstream — AI desk →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →