Only the local model got this word-usage question right — system prompt 'slop' blamed
nomorebuttsplz · reddit · 2026-10-17
A Redditor asked three models (Sonnet 5.5 Medium, GLM 5.3 Flash, GPT 6 Astra Max) whether "The width of the Aptera is a travesty" is correct usage of "travesty." Only the local model answered correctly. The author blames thousands of tokens of sycophantic, safety-maxxed system-prompt instructions injected by frontier labs into web interfaces.
Screenshots show the model first claiming dictionaries record a looser "shockingly bad" sense, then — after pushback — searching and admitting all three major dictionaries tie the definition to imitation/misrepresentation.
The author's argument: this degradation could matter in high-stakes contexts like medical chatbots, and paying for API access or running local models is effectively paying to remove the system prompt that makes models dumber. He concedes it's a single anecdote.
More from Models
- OpenAI's Decisions API hits GA: near real-time routing up to 10x faster than GPT-6 Luna — OpenAIDevs · 2026-10-23
- DeepLearning.AI weekly: GPT-6.1 hits 97.9% byte parity on COBOL migration — DeepLearningAI · 2026-10-23
- Marker's lite mode scores 75.1 vs 42.3 for Azure Read on OmniParseBench — VikParuchuri · 2026-10-23
- Datalab launches lite OCR at $0.50 per 1,000 pages, 20-30x cheaper than Azure, Google and AWS layout APIs — VikParuchuri · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- PHP Foundation reveals Anthropic's Claude cyber model scanned PHP, validated with OpenAI's Daybreak Red — saibharadwaj · 2026-10-23