On Tuesday afternoon, Anthropic launched Claude Performs Pokémon on Twitch, a dwell stream of Anthropic’s latest AI mannequin, Claude 3.7 Sonnet, taking part in a recreation of Pokémon Purple. It’s develop into an enchanting experiment of types, showcasing the capabilities of right now’s AI tech and folks’s reactions to them.
AI researchers have used all types of video video games, from Avenue Fighter to Pictionary, to check new fashions — typically extra for amusement than utility. However Anthropic mentioned that Pokémon proved to be a helpful benchmark for Claude 3.7 Sonnet, which may successfully “assume” via the kinds of puzzles the sport comprises.
Like OpenAI’s o3-mini and DeepSeek’s R1, Claude 3.7 Sonnet can “purpose” its method via robust challenges, like taking part in a online game designed for kids. Whereas the mannequin’s non-reasoning predecessor, Claude 3.5 Sonnet, failed the very starting of Pokémon Purple — exiting the participant’s dwelling in Pallet City — Claude 3.7 Sonnet managed to win three fitness center chief badges.
The most recent Claude nonetheless runs into hassle, although. Hours into the Twitch stream, the mannequin was deterred by a rock wall, which it couldn’t stroll via regardless of how onerous it tried.
One Twitch consumer summed up the state of affairs this manner: “who would win, a pc AI with hundreds of hours put into programming it, or 1 rock wall?”
Finally, Claude realized that it might navigate across the wall.
On the one hand, it’s irritating to observe Claude traverse Pokémon Purple with the pace of a Slowpoke, reasoning via each step with excruciating contemplation. But it’s additionally oddly compelling. The left of the stream exhibits Claude’s “thought course of,” whereas the fitting exhibits real-time gameplay.
At one level, Claude tried to find Professor Oak inside his laboratory, however obtained confused, as a result of there have been different NPCs within the scene.
“I discover a brand new character has appeared under me — a personality with black hair and what seems to be a white coat at coordinates (2, 10),” Claude wrote. “This is likely to be Professor Oak! Let me go down and discuss to him.”
Claude then proceeded to mistakenly discuss to an NPC apart from the Processor — an NPC the mannequin had spoken with a number of instances earlier than. A number of the thousand-odd individuals within the Twitch chat began to get antsy. Others, significantly those that’d been watching the stream for quite a lot of minutes, have been much less nervous.
“Guys chill,” one particular person wrote within the chat. “Earlier than we exited and entered Oak’s lab like 10 instances earlier than understanding find out how to transfer on.”
For longtime Twitch customers, the format of Anthropic’s stream may really feel nostalgic. Over a decade in the past, tens of millions of individuals tried to play Pokémon Purple without delay in a first-of-its-kind on-line social experiment referred to as Twitch Performs Pokémon. Every consumer might management the participant character through Twitch chat, leading to predictably chaotic gameplay.
Some AI researchers have cited Twitch Performs Pokémon as an inspiration for his or her work. In October 2023, Seattle-based software program engineer Peter Whidden printed a YouTube video detailing how he skilled a reinforcement studying algorithm to play Pokémon. His AI spent over 50,000 hours taking part in the sport earlier than it discovered to efficiently navigate it. One problem was that the AI most well-liked to admire the pixelated surroundings as a substitute of really taking part in the sport.
AI-powered “reenactments” of Twitch Performs Pokémon like Whidden’s and Anthropic’s are entertaining, however a bit bittersweet on the identical time. The unique stream was such a pivotal second in Twitch historical past as a result of it introduced individuals collectively in an surprising method. Everybody was on the identical workforce, working towards the objective of getting the participant character to cease operating in circles and truly progress via the sport.
In 2025, it appears we’re not teammates, however spectators, watching an AI mannequin attempt to play a recreation many people obtained the dangle of after we have been 5 years previous. It’s an AI-motivated microcosm of a bigger pattern: our experiences on-line are shifting from shared, communal actions to extra solitary ones.
