{"slug": "gpt-6-2-astra-and-gpt-6-2-sol", "title": "GPT‑6.2 Astra and GPT‑6.2 Sol", "summary": "OpenAI released GPT-6.2 Astra, its new frontier model for hard projects, and GPT-6.2 Sol, an efficient model for everyday work, both available in the API as gpt-6.2-astra and gpt-6.2-sol, according to Simon Willison, who shipped LLM 0.38 with support for the models. GPT-6.2 Astra is priced at $8/MTok input and $40/MTok output — 80% of GPT-6 Astra's $10/$50 and the same as Claude Fable 5.2 — with cached input at $0.40/million versus $1/million for the older model, while GPT-6.2 Sol holds at $2/$10, half the price of GPT-5.6 Sol's $4/$20. Willison reported mixed early reception: Astra 6.2 trails Claude Fable 5.2 on various benchmarks, and Theo Browne's pre-release feedback letter to OpenAI cited the model blocking a two-hour task to ask an unanswerable internal implementation question.", "body_md": "We’re introducing GPT‑6.2 Astra, our new frontier model for your hardest projects, and GPT‑6.2 Sol, our efficient, versatile model for everyday work.\n\nThey're available in the API as `gpt-6.2-astra` and `gpt-6.2-sol`. I shipped [LLM 0.38](https://llm.datasette.io/en/stable/changelog.html#v0-38) with support for the new models.\n\n6.2 Astra is 80% of the price of 6 Astra - $8/MTok input and $40/MTok output, the same price as Claude Fable 5.2 (which was also 80% of its predecessor). 6 Astra and Fable 5 were both $10/$50. The cached price is an even bigger discount - $0.40/million compared to $1/million for the older model. This is great news for longer running agentic sessions.\n\nGPT-6.2 Sol is the same $2/$10 price as 6.1 Sol and 6 Sol - which are half the price of GPT-5.6 Sol ($4/$20).\n\nHere are the pelicans [for GPT-6.2 Sol](https://tools.simonwillison.net/markdown-svg-renderer?url=https%3A%2F%2Fgist.github.com%2Fsimonw%2Fa940cc2750162ac9edf0ddf881067b83) and [for GPT-6.2 Astra](https://tools.simonwillison.net/markdown-svg-renderer?url=https%3A%2F%2Fgist.github.com%2Fsimonw%2Ffe836e13c90c00233c9e9e499a941805). The 6.2 Astra max pelican looks like this:\n\nI have a [full comparison grid here](https://static.simonwillison.net/static/2026/openai-pelicans-november-2026.html), but here's how the high/xhigh/max pelicans compare with the previous Sol and Astra releases:\n\nI also ran my most recent benchmark prompt against the two models, using \"Ultra' effort levels in the Codex CLI:\n\n`Make a three js animation where yonkers stands up as a big mech (cool stand up motion) and stomps around manhattan and queens and Brooklyn, use openstreetmap mapping data`\n\nHere's [Sol](https://static.simonwillison.net/static/2026/yonkers-gpt-6.2-sol-ultra/index.html) and here's [Astra](https://static.simonwillison.net/static/2026/yonkers-gpt-6.2-astra-ultra/index.html). Neither of them did a good enough job of morphing Yonkers into a big mech, at least in comparison to [Claude Fable 5.2](https://tools.simonwillison.net/yonkers-rises) and [Gemini 4 Carbon](https://static.simonwillison.net/static/2026/yonkers-mech-standalone.html).\n\nThis video showing the Astra and Sol animations playing out at the same time:\n\nI created that video using GPT-6.2 Astra (Ultra) in the Codex desktop app by prompting:\n\n`https://static.simonwillison.net/static/2026/yonkers-gpt-6.2-astra-ultra/index.html and https://static.simonwillison.net/static/2026/yonkers-gpt-6.2-sol-ultra/index.html  are two videos of the same thing by two different models`\n\n`I need a video of the first 10 seconds where the top half of the video plays the first and the bottom half plays the second`\n\n`I need an overlaid label of \"GPT-6.2 Astra\" on the top half and \"GPT-6.2 Sol\" on the bottom half. These labels should be tastefully positioned.`\n\n`Make the video square and 2000x2000 pixels`\n\nThen later:\n\n`OK make it 20s`\n\nHere's the [exported transcript](https://gist.github.com/simonw/6b44c588a1321b60799dbbfd13db6d23).\n\nThe buzz around these new models is mixed at the moment. Astra 6.2 isn't scoring quite as high as Fable 5.2 on various benchmarks, and I've seen a few complaints that its communication style can be confusing. Theo Browne's [review](https://twitter.com/theo/status/2122760447103484410) (in the form of a feedback letter he sent OpenAI during pre-release) has this useful example:\n\nWhen this model is working on long tasks, it can pull in a lot of subagents and manage them pretty well. It will then proceed to randomly *ask you some weird internal implementation detail bullshit that you could not under any earthly circumstance understand, much less have input on*.\n\nI sent \"/goal make a new level for Outer Wilds\". 2 hours in, it's \"blocked\" because it wanted to ask me \"does r24's reviewer write it's clips to 24_g521.mp4 or 24_p11.mp4?\"\n\nTags: [ai](https://simonwillison.net/tags/ai), [openai](https://simonwillison.net/tags/openai), [generative-ai](https://simonwillison.net/tags/generative-ai), [llms](https://simonwillison.net/tags/llms), [llm-pricing](https://simonwillison.net/tags/llm-pricing), [pelican-riding-a-bicycle](https://simonwillison.net/tags/pelican-riding-a-bicycle), [llm-release](https://simonwillison.net/tags/llm-release), [theo-browne](https://simonwillison.net/tags/theo-browne), [gpt](https://simonwillison.net/tags/gpt), [yonkers-rising](https://simonwillison.net/tags/yonkers-rising)", "url": "https://wpnews.pro/news/gpt-6-2-astra-and-gpt-6-2-sol", "canonical_source": "https://simonwillison.net/2026/Nov/17/gpt62-astra-and-sol/", "published_at": "2026-11-17 22:38:52+00:00", "updated_at": "2026-11-18 15:15:54.950574+00:00", "lang": "en", "topics": ["large-language-models", "generative-ai", "ai-products", "ai-tools", "ai-agents"], "entities": ["OpenAI", "GPT-6.2 Astra", "GPT-6.2 Sol", "Simon Willison", "LLM 0.38", "Claude Fable 5.2", "Gemini 4 Carbon", "Theo Browne"], "also_reported_by": [], "takeaway": "GPT-6.2 Astra costs $8/MTok input and $40/MTok output with cached input at $0.40/million — 80% of GPT-6 Astra's $10/$50 and a 60% cut on cached tokens — while GPT-6.2 Sol stays at $2/$10, and both are callable as gpt-6.2-astra and gpt-6.2-sol in the API and via LLM 0.38.", "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/gpt-6-2-astra-and-gpt-6-2-sol", "markdown": "https://wpnews.pro/news/gpt-6-2-astra-and-gpt-6-2-sol.md", "text": "https://wpnews.pro/news/gpt-6-2-astra-and-gpt-6-2-sol.txt", "jsonld": "https://wpnews.pro/news/gpt-6-2-astra-and-gpt-6-2-sol.jsonld"}}