Dev says fine-tuned open-weight models beat closed ones at deep research
gnukeith · x · 2026-11-22
Developer gnukeith pushes back on the idea that open-weight models see no real use: he uses Opus for personal coding, but claims his fine-tuned research models significantly outperform closed models—especially Anthropic's—at deep research. He jailbreaks TVs with DeepSeek and secures environments with GLM plus the mimosa plugin, arguing closed labs' safety focus tanks practical usefulness.
Related event: Debates flare over whether open models are dead(2 posts)→
More from Models
- User Reports Noticeable Quality Drop in Claude Opus 5.5 Recently — jiayuan_jy · 2026-11-22
- Speculation: OpenAI quietly throwing massive compute at remaining Millennium Problems — haider1 · 2026-11-22
- Fine-tuned 2B models beat OpenAI o3 and Gemini 2.5 few-shot on molecular classification — madeofAjala · 2026-11-22
- Sakana AI announces its own pretrained LLM series Sakana Sawara, first model due spring 2027 — SakanaAILabs · 2026-11-22
- HF's harness adapters lift Qwen3.8-27B from 39% to 55% on Terminal-Bench 2.1 with half the tokens — SergioPaniego · 2026-11-22
- KOL: a month on, none of China's four flagship models comes close to US frontier — kimmonismus · 2026-11-22