{"slug": "how-i-view-llms-as-sept-2026", "title": "How I view LLMs as Sept 2026", "summary": "A developer who has spent over $30,000 on tokens across Codex, Claude, and GLM since the start of the year argues that large language models now fall into two categories: instruction-following models like Luna and Sonnet, which are cheap workhorses that lack common-sense judgment, and intent-understanding frontier models like Astra and Fable, which better grasp the intent behind a question and can orchestrate the cheaper models. The developer contends that today's models remain far from making human-like trade-offs, and that closing that gap is the path to AGI.", "body_md": "### How I view LLMs as Sept 2026\n\nI’ve spent over $30k in tokens since the start of the year across Codex, Claude, and GLM.\n\nHere’s how I view LLMs as of September 2026. Two categories:\n\n**Instructions-following models:**\n\n- Like Luna and Sonnet.\n- I used to call these “dumb”. But that’s the wrong way to think about them.\n- They just cannot make good “common sense” decisions.\n- They’re workhorses for pennies. Luna especially, is hard to spend a lot on.\n- If you make the core decisions up front, coding is auto-complete.\n\n**Intent-understanding models:**\n\n- Frontiers like Astra and Fable.\n- These are higher “common sense” models. These models (try to) understand the intent behind the question.\n- Solving problems is about making the right tradeoff - these models tend to make the right tradeoff more often than the other models.\n- These models know how to orchestrate instruction-following models prescriptively - in a way humans get too lazy to.\n\nIn my experience, today’s models are far from making the right human-like trade-offs.\n\nHowever, this is where I believe the path to AGI is.\n\nThat is, if a model can make the tradeoffs a human would.", "url": "https://wpnews.pro/news/how-i-view-llms-as-sept-2026", "canonical_source": "https://bhardwajrish.blogspot.com/2026/09/how-i-view-llms-as-sept-2026.html", "published_at": "2026-09-21 00:04:09+00:00", "updated_at": "2026-09-21 00:23:18.585765+00:00", "lang": "en", "topics": ["large-language-models", "artificial-intelligence", "ai-agents", "ai-research"], "entities": ["Codex", "Claude", "GLM", "Luna", "Sonnet", "Astra", "Fable"], "alternates": {"html": "https://wpnews.pro/news/how-i-view-llms-as-sept-2026", "markdown": "https://wpnews.pro/news/how-i-view-llms-as-sept-2026.md", "text": "https://wpnews.pro/news/how-i-view-llms-as-sept-2026.txt", "jsonld": "https://wpnews.pro/news/how-i-view-llms-as-sept-2026.jsonld"}}