Home/Blog/Open models can do 90% of work at a twentieth of the cost of a frontier LLMAI & MLOpen models can do 90% of work at a twentieth of the cost of a frontier LLMHacker News New2026-11-03 23:410 viewsArticle URL: https://www.theregister.com/ai-and-ml/2026/11/03/open-models-can-do-90-of-agentic-work-at-a-twentieth-of-the-cost-of-a-frontier-llm-gartner/5307132 Comments URL: https://news.ycombinator.com/item?id=50353385 Points: 1 # Comments: 0Source:https://www.theregister.com/ai-and-ml/2026/11/03/open-models-can-do-90-of-agentic-work-at-a-twentieth-of-the-cost-of-a-frontier-llm-gartner/5307132My favoritesPreviousShow HN: HackerChurn – which SaaS tools HN wants to cancelNextBear turns 10, and it still feels like day oneRelated postsWhy 80% on OSWorld Required VM Snapshot RollbacksIf you have ever watched an autonomous computer-use agent work through a desktop operating system, you know the exact mo0dev.to AILLM API Pricing Compared: Claude vs GPT vs Gemini vs DeepSeek (Live Numbers, Nov 2026)Short version: the gap between a model vendor's list price and what you actually pay on a multi-seller LLM gateway can b0dev.to AIVoice Jitter Buffer: The Hardest Part of Building AI Voice AgentsWhen people talk about AI voice agents, they usually talk about the model. They talk about prompts, tools, and how smart0dev.to AIHow to Do a Reverse Face Search on a Dating Profile Photo (Step-by-Step, 2026)To reverse face search a dating photo, crop the face, then run it through a face search engine. ProfileFinder.ai checks 0dev.to SecurityHow to fine-tune an embedding model with LoRA adaptersTo fine-tune an embedding model with LoRA, freeze the base encoder, attach a small LoRA adapter to its linear layers, an0dev.to PythonCómo saber el canal real de cada lead en GoHighLevel (y por qué utmSessionSource te engaña)Si trabajas con GoHighLevel (GHL) y alguna vez has intentado responder a la pregunta «¿de dónde vienen mis leads?», prob0dev.to Python