By using this website, you agree to our Privacy Policy and Terms of Use.
Accept
Craftium.AICraftium.AICraftium.AI
  • Home
  • News
  • Catalog
  • Collections
  • Blog
Font ResizerAa
Craftium.AICraftium.AI
Font ResizerAa
Пошук
  • Home
  • News
  • Catalog
  • Collections
  • Blog
Follow US
  • Terms of Use
  • Privacy Policy
  • Copyright
  • Feedback
© 2024-2025 Craftium.AI.

Claude 3.7 Sonnet Successfully Plays Pokémon

Anthropic tests AI capable of defeating gym leaders in the classic Game Boy game

Eleni Karasidi
Eleni Karasidi
Published: 25.02.2025
News
Claude Sonnet
Image from Anthropic’s blog post
SHARE

Anthropic has tested its new generative AI model, Claude 3.7 Sonnet, using the classic Game Boy game — Pokémon Red. In its blog, Anthropic reported that the model was equipped with basic memory, could receive input from screen pixels, and perform function calls to press buttons and navigate the screen, which allowed it to play Pokémon continuously.

Read also

AWS
AWS in partnership with Anthropic launches AI agents marketplace
Apple Tests OpenAI and Anthropic Models for Next-Generation Siri
AI Assistant Claude Managed Anthropic Store for a Month

One of the unique features of Claude 3.7 Sonnet is its ability for “enhanced reasoning.” This enables the model to solve complex tasks by applying more computational resources and spending more time. Compared to the previous version, Claude 3.0 Sonnet, which was unable to leave the house in Pallet Town, Claude 3.7 Sonnet successfully defeated three gym leaders in the game and earned their badges.

Although it is still unknown how many computational resources were required to achieve these results, the company noted that the model performed 35,000 actions to reach the last gym leader named Surge.

Claude allows users to create AI applications without programming
US Court Allows Anthropic to Train AI on Purchased Books
Leading AI Models Exhibit Harmful Autonomy in Anthropic Tests
AI is taught spatial thinking through Snake and Tetris games
AI Models Get Confused While Playing Pokémon
TAGGED:AnthropicGamingTesting
Leave a Comment

Leave a Reply Cancel reply

Follow us

XFollow
YoutubeSubscribe
TelegramFollow
MediumFollow

Popular News

Mistral AI
Mistral offers the open language model Voxtral for speech
17.07.2025
Grok 4
Grok 4 from xAI promises to become the new standard for developers
06.07.2025
ChatGPT Agent
OpenAI introduced ChatGPT Agent, allowing AI to delegate complex tasks
18.07.2025
Chatbots in messengers
Meta chatbots will gain the ability to independently send messages to users
04.07.2025
FLUX.1 context [dev]
FLUX.1 context [dev] simplifies image generation for all users
02.07.2025

Читайте також

resisting robot
News

Research Reveals GPT-4o’s Reluctance to Shut Down

13.06.2025
Reddit
News

Reddit Challenges Anthropic’s Actions Over Content Access

04.06.2025
windsurf
News

Windsurf Faces Sudden Access Restrictions to Claude

04.06.2025

Craftium AI is a team that closely follows the development of generative AI, applies it in their creative work, and eagerly shares their own discoveries.

Navigation

  • News
  • Reviews
  • Collections
  • Blog

Useful

  • Terms of Use
  • Privacy Policy
  • Copyright
  • Feedback

Subscribe for AI news, tips, and guides to ignite creativity and enhance productivity.

By subscribing, you accept our Privacy Policy and Terms of Use.

Craftium.AICraftium.AI
Follow US
© 2024-2025 Craftium.AI
Subscribe
Level Up with AI!
Get inspired with impactful news, smart tips and creative guides delivered directly to your inbox.

By subscribing, you accept our Privacy Policy and Terms of Use.

Welcome Back!

Sign in to your account

Username or Email Address
Password

Lost your password?