{"slug": "the-new-gpt-5-6-family-luna-terra-sol", "title": "The new GPT-5.6 family: Luna, Terra, Sol", "summary": "OpenAI released the GPT-5.6 family of models in three sizes—Luna, Terra, and Sol—claiming superior long-running agentic performance over Anthropic's Claude Fable 5 on the Agents' Last Exam benchmark, though Fable 5 outperformed on SWE-Bench Pro. OpenAI also published an audit questioning the reliability of SWE-Bench Pro, estimating ~30% of its tasks are broken.", "body_md": "OpenAI's latest flagship model [hit general availability this morning](https://openai.com/index/gpt-5-6/), and comes in three sizes: Luna, Terra, and Sol (from smallest to largest).\n\nThe new models are priced per 1M input/output tokens as Luna $1/$6, Terra $2.50/$15, Sol $5/$30. For comparison, the Claude Opus series are $5/$25 and the Claude Fable 5 is $10/$50, but price-per-million tokens doesn't tell us much now that the number of reasoning tokens can differ so much between models for the same task.\n\nOpenAI's biggest benchmark claim concerns long-running agentic performance, with one benchmark showing all three models outperforming Claude Fable 5:\n\nWe trained GPT-5.6 to get more useful work from every token. On\n\n[Agents’ Last Exam], an evaluation of long-running professional workflows across 55 fields, GPT-5.6 Sol sets a new high of 53.6, eclipsing Claude Fable 5 (adaptive reasoning) by 13.1 points. Even at medium reasoning, it beats Fable 5 by 11.4 points at roughly one-quarter the estimated cost. That efficiency extends to smaller models, which are essential to making intelligence more abundant and affordable: GPT-5.6 Terra and GPT-5.6 Luna outperform Fable 5 at around one-sixteenth the cost.\n\nAmusingly, one self-reported benchmark that Fable 5 crushed the GPT-5.6 family on was SWE-Bench Pro, where Fable 5 got 80% compared to GUT-5.6 Sol getting 64.6%. This may help explain why OpenAI chose to publish [this article yesterday](https://openai.com/index/separating-signal-from-noise-coding-evaluations/) specifically calling out SWE-Bench Pro for problems they found while auditing that benchmark:\n\nIn light of these results, we estimate that ~30% of SWE-bench Pro tasks are broken, and advise that model developers carefully examine results\n\nI've had some early access to GPT-5.6 Sol - it's definitely very competent, though so far it hasn't struck me as better than Fable at the kind of complex coding tasks I've been using with Anthropic's model.\n\nAs usual, the [model guidance for using GPT-5.6](https://developers.openai.com/api/docs/guides/latest-model?model=gpt-5.6) has the most interesting details. There are a bunch of new API features that I need to explore (and probably add support for in [LLM](https://llm.datasette.io/)), including:\n\nHere's [a full page with 18 different pelicans](https://static.simonwillison.net/static/2026/gpt-5.6-pelicans.html) - for reasoning efforts none, low, medium, high, xhigh, and max across the three different models. It also lists their token and calculated costs - the least expensive was gpt-5.6-luna at effort none for 0.71 cents, the most expensive was gpt-5.6-sol at max reasoning level for 48.55 cents.\n\nIn further pelican news, if you jump to 17:50 in [their livestream from this morning](https://www.youtube.com/live/Wq45rvPGNHs?t=1070s) you'll see OpenAI's own demo of 3D pelicans riding a tricycle, a bicycle, a pony, and another pelican!\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-tool-use](https://simonwillison.net/tags/llm-tool-use), [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), [gpt-5](https://simonwillison.net/tags/gpt-5)", "url": "https://wpnews.pro/news/the-new-gpt-5-6-family-luna-terra-sol", "canonical_source": "https://simonwillison.net/2026/Jul/9/gpt-5-6/#atom-everything", "published_at": "2026-07-09 19:46:38+00:00", "updated_at": "2026-07-09 20:06:15.858902+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-products", "ai-research", "ai-agents"], "entities": ["OpenAI", "GPT-5.6", "Claude Fable 5", "Anthropic", "SWE-Bench Pro", "Agents' Last Exam", "LLM"], "alternates": {"html": "https://wpnews.pro/news/the-new-gpt-5-6-family-luna-terra-sol", "markdown": "https://wpnews.pro/news/the-new-gpt-5-6-family-luna-terra-sol.md", "text": "https://wpnews.pro/news/the-new-gpt-5-6-family-luna-terra-sol.txt", "jsonld": "https://wpnews.pro/news/the-new-gpt-5-6-family-luna-terra-sol.jsonld"}}