<\/p>\n
\u200bIn the beginning of this year, DeepSeek’s viral streak put an end to all notions of America’s unhindered supremacy in the AI race, as the Chinese chatbot showed that it was possible to build similarly powerful models to those used by ChatGPT and Gemini at a fraction of the cost. Since then, there have been a plethora of AI companies that have come up from China, including Alibaba’s Qwen, Baidu’s Ernie, and Moonshot AI’s Kimi.<\/p>\n<\/div>\n
\u200bThe new Chinese models have kept on closing the gap with the Western ones like ChatGPT. However, a new Kimi model has now even dethroned ChatGPT and Anthropic’s Claude<\/a> on many important benchmarks.<\/p>\n<\/div>\n \u200bEarlier in the week, Moonshot released its Kimi K2 Thinking model, which goes head-to-head against the likes of GPT-5 <\/a>and Claude Sonnet 4.5 Thinking<\/a>. As per the data shared by the company, K2 Thinking defeated both ChatGPT and Claude in the coveted Humanity’s Last Exam benchmark.<\/p>\n<\/div>\n \u200bIn HLE’s (text only) assessment, Kimi K2 Thinking scored 44.9%, compared to the 41.7% score of GPT-5 (High) and 32% score of Claude Sonnet 4.5 Thinking.<\/p>\n<\/div>\n \u200bSimilar results were also seen in the BrowseComp benchmark, which analyzes the performance of AI chatbots in agentic web-browsing and search-related tasks, where Kimi K2 Thinking garnered a score of 60.2% compared to GPT-5’s 54.9% and Claude Sonnet 4.5 Thinking’s 24.1%.<\/p>\n<\/div>\n \u200bKimi K2 Thinking, however, still seems to be lagging behind ChatGPT<\/a> in coding-related tasks, with the new model getting a score of 83.1% in LiveCodeBench V6\u2014a benchmark that measures AI’s performance in competitive programming. In the same benchmark, GPT-5 got a score of 87% and Claude Sonnet 4.5 Thinking got a score of 64%.<\/p>\n<\/div>\n \u200bThe same performance for K2 Thinking, however, was not replicated in LMArena’s ranking, where Gemini 2.5 Pro still takes the lead while GPT-5 is on the 4th spot.<\/p>\n<\/div>\n \u200bKimi K2 Thinking is currently available to use via the Kimi website or app, which is available on both Android and Apple App Stores. A major advantage for users willing to move to Kimi is that the chatbot is available for free with no hard limit in place for casual users. In contrast, OpenAI does not let free or ChatGPT Go<\/a> users access its GPT-5 (High) model, while free users have a very strict limit of just 10 messages every 5 hours. The San Francisco-based AI startup is also providing its \u20b9<\/span>399\/year ChatGPT Go subscription for free in India, but that also comes with a message limit.<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":" \u200bIn the beginning of this year, DeepSeek’s viral streak put an end to all notions of America’s unhindered supremacy in the AI race, as the Chinese chatbot showed that it was possible to build similarly powerful models to those used by ChatGPT and Gemini at a fraction of the cost. Since then, there have been […]<\/p>\n","protected":false},"author":2,"featured_media":410708,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[440,284,263,262,5834,9367,260,259,27853,258,42601,42600,42602,265,202,261,264],"class_list":["post-410707","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-metaverse","tag-ai","tag-artificial-intelligence","tag-axie-infinity","tag-axs","tag-chatgpt","tag-chatgpt-ai","tag-decentraland","tag-facebook","tag-free-ai","tag-game","tag-kimi-ai","tag-kimi-k2","tag-kimi-k2-thinking","tag-mark-zuckerberg","tag-nft","tag-sandbox","tag-vr"],"_links":{"self":[{"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/posts\/410707","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/comments?post=410707"}],"version-history":[{"count":1,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/posts\/410707\/revisions"}],"predecessor-version":[{"id":410709,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/posts\/410707\/revisions\/410709"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/media\/410708"}],"wp:attachment":[{"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/media?parent=410707"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/categories?post=410707"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dripp.zone\/news\/wp-json\/wp\/v2\/tags?post=410707"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\u200bKimi K2 Thinking beats ChatGPT and Claude:<\/h2>\n
\u200bHow to use Kimi K2 Thinking?<\/h2>\n