{"slug": "gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices", "title": "GPT-5.6 Sol's Factorio rocket run used 250 million tokens, about $52 at API prices, tester says", "summary": "Tester Ryan Madden reported on Oct. 31 that GPT-5.6 Sol completed a peaceful Factorio run, launching a rocket through a Python interface, at a cost of about 250 million input tokens (95% cached) and 1.5 million output tokens — roughly $52 at API pricing — over 13.5 hours of game time. Madden said the model struggled to connect the factory's late-game production systems, and his token breakdown offers a qualified cost reference for long-running game tasks.", "body_md": "# GPT-5.6 Sol's Factorio rocket run used 250 million tokens, about $52 at API prices, tester says\n\n**Madden said the run used about 250 million input tokens, 95% cached, and 1.5 million output tokens.**\n\n        By [Emerson Redding](https://runtimewire.com/author/emerson-redding)\n        · Published \n\nPrimary source: [ryanmadden.net](https://ryanmadden.net/gpt-56-beats-factorio)\n\n## Why it matters\n\nMadden's token breakdown offers a qualified cost reference for a long-running game task.\n\nRyan Madden reported Oct. 31st that GPT-5.6 Sol completed a peaceful Factorio run, using a Python interface to launch a rocket despite struggling to connect the factory's late-game production systems.\n\nMadden told RuntimeWire in an email: “It was about 250M input tokens (95% cached) and 1.5M output which would have been $52 at API pricing. The run took 13.5 hours of game time.”…\n\nWhen you purchase through links in our articles, [we may earn a small commission](https://runtimewire.com/affiliate-monetization-standards). This doesn’t affect our editorial independence.", "url": "https://wpnews.pro/news/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices", "canonical_source": "https://runtimewire.com/article/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices", "published_at": "2026-10-31 21:20:54+00:00", "updated_at": "2026-11-01 00:50:45.802658+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-agents"], "entities": ["GPT-5.6 Sol", "Ryan Madden", "Factorio", "RuntimeWire"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices", "markdown": "https://wpnews.pro/news/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices.md", "text": "https://wpnews.pro/news/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices.txt", "jsonld": "https://wpnews.pro/news/gpt-5-6-sol-s-factorio-rocket-run-used-250-million-tokens-about-52-at-api-prices.jsonld"}}