loading
loading
OpenAI’s GPT-6 Astra model autonomously completes Portal in 24 hours — feat cost just $571 in tokens Astra had to map 3D game spaces and understand the puzzles before it could plan solutions and succeed. An AI and LLM enthusiast has conducted an experiment where OpenAI's new GPT-6 Astra played through the entirety of Valve's Portal 3D puzzler game on its own. This seems like a remarkably progressive step forward for LLMs and a convincing demonstration of multimodal 'AI intelligence.' It wasn’t that long ago AIs were losing at Atari 2600 chess. However, this puzzle gaming task resulted in 3,336 tool calls and a headline API cost of $571.18. CozyBlaze, the enthusiast, has since clarified that the costs were covered by their $200 Codex Pro subscription. The video above has Astra’s thinking/pauses removed to make watching somewhat bearable. “The model controls Portal through MCP [Model Context Protocol] + a modified SourcePauseTool,” explains CozyBlaze. “The game stays paused while the model thinks; once the model sends an input sequence, SPT unpauses and executes it.” During the thinking time, the AI received screenshots and information about the player character position. After this, the game resumed, and Astra executed its planned moves and inputs. This is why the edited highlights reel is ~ 2 hours, but the full set of GPT-6 Astra Portal VOD streams adds up to ~24 hours. CozyBlaze is pragmatic about the achievement. They admit there are still many problems to solve, and this runthrough shouldn’t be viewed as an AI benchmark. “But watching a general-purpose agent autonomously navigate and make it all the way through the game feels like a small glimpse of that original vision becoming real,” they observe. That original vision is OpenAI’s from way back in 2016, when an eventual goal of a single agent solving a variety of games was mentioned. If you want to see the inner workings of CozyBlaze’s Portal Agent, it is available on GitHub. There are some instructions there for you to explore it or try it for yourself, too. GPT-6 Astra became OpenAI’s flagship model earlier this month. OpenAI claims the new model offers “a new generation of intelligence,” and “is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work.” Expect to see more of Astra flexing its muscular gray matter in the coming days. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason. - usertests It takes many tens of thousands of dollars (depending on standard of living) to raise a baby to adulthood, so this is a good result.Reply - USAFRet Reply Why is this a 'good result'?usertests said:It takes many tens of thousands of dollars (depending on standard of living) to raise a baby to adulthood, so this is a good result. Are humans creating games just for an AI entity to solve them? - Dementoss This is still far from true intelligence and, just an AI agent executing a long list of instructions.Reply - n3rdi When asked for comment after completing the game, the AI simply asked, "When is the sequel going to be released?"Reply - alrighty_then That's a nice achievement. How long till someone opens a Steam account for their AI and has it beat every game and get every reward, just because? Less than years away...Reply - killerbishop So, did the LLM have access to the Portal strategy guide, images of how to solve the puzzles, or any other data that it had previous training from? If so, this becomes less convincing that the AI solved the puzzle solely by observation. The use of MCP commands might help it know what commands to input, but if it just map those to gamefaqs, I'm not sure this is that much of a leap in novel learning capabilities.Reply - Fox2Fox2 Reply If you're going there, you have to include the billions spent training the model.usertests said:It takes many tens of thousands of dollars (depending on standard of living) to raise a baby to adulthood, so this is a good result. - Fox2Fox2 Reply Yes, undeniably it would have. The likelihood of multiple online guides for a game as popular as Portal being in the models dataset is almost guaranteed, at least for a model the size of Astra. It's pretty much unavoidable.killerbishop said:So, did the LLM have access to the Portal strategy guide, images of how to solve the puzzles, or any other data that it had previous training from? If so, this becomes less convincing that the AI solved the puzzle solely by observation. The use of MCP commands might help it know what commands to input, but if it just map those to gamefaqs, I'm not sure this is that much of a leap in novel learning capabilities.
document fetched 2026-09-08 02:57 · canonical https://tomshardware.com/tech-industry/artificial-intelligence/openais-gpt-6-astra-model-autonomously-completes-portal-in-24-hours-feat-cost-just-usd571-in-tokens