Users Report Claude's New Models Lag Behind Opus 4.6
anmarasovic · x · 2026-07-04
Researchers note that early Claude models like Opus 4.6 could accurately infer user intent and provide concise, non-redundant answers. In contrast, while Fable 5 and Claude 4.8 possess greater knowledge, their ability to understand "what users truly want" has noticeably deteriorated. The user's curiosity about the technical reasons behind this regression sparked a discussion on the degradation of instruction-following capabilities in new models.
More from Models
- GPT-6 Astra beats Factorio with enemies in 44 in-game hours at ~$4,500 API cost — liminal_bardo · 2026-09-11
- giffmana: the env being used in training is part of the point — giffmana · 2026-09-11
- awesome-llm-leaderboards: an open-source directory of LLM leaderboards, pricing tables, comparison tools — Last_Establishment_1 · 2026-09-11
- Anthropic claims it works to keep eval environments unidentifiable to models — MaxKannen · 2026-09-11
- Nex N2.5 Pro released on Hugging Face with 407GB of weights — jinnyjuice · 2026-09-11
- RoMa v2 image matching model unveiled in the usual black poster — ducha_aiki · 2026-09-11