GPT-6 Astra Autonomously Completes Portal Game in 24 Hours, Costs $571 in API Calls

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.

GPT-6 Astra just completed Portal, but it burned through $571.18 in tokens to do it

* GPT-6 Astra autonomously beat Portal, showing big LLM gameplay progress. * Gameplay looked aimbot-like; video edits removed pauses while the model 'thought.' * The run cost roughly $571.18 in API calls -- an impressive but expensive proof of progress. About a year and a half ago, I watched Claude try to beat a Pokémon game on the Twitch channel ClaudePlaysPokemon. Back then, Claude kept getting stuck on things, and it was mostly frustrating to watch him bumble around and undo progress. Now, GPT-6 Astra has shown how AI has evolved by beating Portal by itself, although it had to churn through $571.18 in API calls to do it. GPT-6 Astra manages to beat Portal and leaves its operator with the bill Quite a hefty bill, at that As spotted by VideoCardz, X user cozyblazex made a post announcing that they had watched GPT-6 Astra complete the hit Valve video gamePortal. If you're not sure what that is, GPT-6 Astra is OpenAI's latest, most powerful model yet, and the one the company hopes will bring the world of LLMs into the AGI era. I'm not sure about the AGI part, but at the very least, you can see it beat Portal in the video above. The motions are a bit jarring because it acts with perfect speed and accuracy, making it look like it's using an aimbot. Well, it's more like it is the aimbot. At any rate, the LLM didn't finish the game in 2 hours, as the video claims; the game was paused every time GPT-6 Astra had to think, and those bits had been cut out of the final video. Unfortunately, the venture did eat into the API calls quite a bit. Beating Portal cost cozyblazex a hefty $571.18 in tokens, which is probably not the most cost-effective way to beat a game. Regardless, it's a very cool feat, and it'll be interesting to see how LLMs evolve to play games in another year or two's time.