{"id":6824,"date":"2025-02-24T20:00:26","date_gmt":"2025-02-24T20:00:26","guid":{"rendered":"https:\/\/fusionmindlabs.com\/blogs\/anthropic-uses-pokemon-to-test-ai-model\/"},"modified":"2025-02-24T20:00:26","modified_gmt":"2025-02-24T20:00:26","slug":"anthropic-uses-pokemon-to-test-ai-model","status":"publish","type":"post","link":"https:\/\/fusionmindlabs.com\/blogs\/anthropic-uses-pokemon-to-test-ai-model\/","title":{"rendered":"Anthropic uses Pok\u00e9mon to test AI model"},"content":{"rendered":"\n<\/p>\n<div>\n<p id=\"speakable-summary\" class=\"wp-block-paragraph\">In a rather unexpected move, Anthropic utilized the popular game Pok\u00e9mon to test its latest AI model, demonstrating its capabilities in a unique way.<\/p>\n<p class=\"wp-block-paragraph\">According to a blog <a rel=\"nofollow noopener\" href=\"https:\/\/www.anthropic.com\/news\/claude-3-7-sonnet\" target=\"_blank\">post<\/a> released on Monday, Anthropic&#8217;s newest model, <a href=\"https:\/\/techcrunch.com\/2025\/02\/24\/anthropic-launches-a-new-ai-model-that-thinks-as-long-as-you-want\/\" rel=\"nofollow noopener\" target=\"_blank\">Claude 3.7 Sonnet<\/a>, was put to the test on the classic Game Boy game Pok\u00e9mon Red. The model was equipped with basic memory, screen pixel input, and function calls to press buttons and navigate the screen, enabling it to play Pok\u00e9mon continuously.<\/p>\n<p class=\"wp-block-paragraph\">One of the distinctive features of Claude 3.7 Sonnet is its ability to engage in &#8220;extended thinking,&#8221; allowing it to tackle complex problems by applying more computing power and taking more time, similar to models like OpenAI&#8217;s o3-mini and DeepSeek&#8217;s R1.<\/p>\n<p class=\"wp-block-paragraph\">This capability proved to be particularly useful in Pok\u00e9mon Red, as evidenced by the model&#8217;s performance.<\/p>\n<p class=\"wp-block-paragraph\">In comparison to its predecessor, Claude 3.0 Sonnet, which was unable to progress beyond the starting house in Pallet Town, Claude 3.7 Sonnet successfully battled three Pok\u00e9mon gym leaders and earned their respective badges.<\/p>\n<figure class=\"wp-block-image aligncenter size-large\"><img loading=\"lazy\" loading=\"lazy\" decoding=\"async\" width=\"1920\" height=\"1269\" src=\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?w=680\" alt=\"Anthropic Pokemon Red\" class=\"wp-image-2970257\" srcset=\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp 1920w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=150,99 150w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=300,198 300w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=768,508 768w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=680,449 680w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=1200,793 1200w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=1280,846 1280w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=430,284 430w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=720,476 720w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=900,595 900w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=800,529 800w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=1536,1015 1536w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=668,442 668w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=567,375 567w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=934,617 934w, https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/02\/9a30c8288e402eee24d4ef60272cb6365a36207a-1920x1269-1.webp?resize=708,468 708w\" sizes=\"auto, (max-width: 1920px) 100vw, 1920px\"\/><figcaption class=\"wp-element-caption\"><span class=\"wp-block-image__credits\"><strong>Image Credits:<\/strong>Anthropic<\/span><\/figcaption><\/figure>\n<p class=\"wp-block-paragraph\">Although the exact amount of computing required for Claude 3.7 Sonnet to achieve these milestones is unclear, as well as the time it took to reach each one, Anthropic did reveal that the model performed 35,000 actions to reach the last gym leader, Surge.<\/p>\n<p class=\"wp-block-paragraph\">It is likely that this information will be uncovered by a developer in the near future.<\/p>\n<p class=\"wp-block-paragraph\">While Pok\u00e9mon Red serves as more of a toy benchmark, there is a long history of using games for AI benchmarking purposes, with several new apps and platforms emerging in recent months to test models&#8217; game-playing abilities on various titles, including <a rel=\"nofollow noopener\" href=\"https:\/\/github.com\/OpenGenerativeAI\/llm-colosseum\" target=\"_blank\">Street Fighter<\/a> and <a href=\"https:\/\/techcrunch.com\/2024\/11\/05\/people-are-using-games-like-pictionary-to-benchmark-ai-now\/\" rel=\"nofollow noopener\" target=\"_blank\">Pictionary<\/a>.<\/p>\n<\/div>\n\n<br \/><a href=\"https:\/\/techcrunch.com\/2025\/02\/24\/anthropic-used-pokemon-to-benchmark-its-newest-ai-model\/\" rel=\"nofollow noopener\" target=\"_blank\">Source Link<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>In a rather unexpected move, Anthropic utilized the popular game Pok\u00e9mon to test its latest AI model, demonstrating its capabilities in a unique way. According to a blog post released&#8230;<\/p>\n","protected":false},"author":0,"featured_media":6825,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[47],"tags":[460,924,2150,1994,2151],"class_list":["post-6824","post","type-post","status-publish","format-standard","has-post-thumbnail","category-news","tag-anthropic","tag-benchmark","tag-claude-3-7-sonnet","tag-pokemon","tag-pokemon-red"],"jetpack_featured_media_url":"https:\/\/fusionmindlabs.com\/blogs\/wp-content\/uploads\/sites\/9\/2025\/02\/pokemon.png","_links":{"self":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts\/6824","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/comments?post=6824"}],"version-history":[{"count":0,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts\/6824\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/media\/6825"}],"wp:attachment":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/media?parent=6824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/categories?post=6824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/tags?post=6824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}